emulsiones de alimentos

tood tmulsions Fourth Edition, Revised and Expanded

edited by Stig E. Friberg h i v e rsity of Missouri- Rolla Rolla, Missouri and Clarkson University Potsdam, New York, U.S.A.

KIre Larsson Camurus Lipid Research Lund, Sweden

Johan Sjoblom Norwegian University of Science and Technology Trondheim, Norway

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

The previous edition of this book was published as Food Emulsions: Third Edition,

Revised and Expanded, edited by Stig E. Friberg and Kare Larsson, 1997 (Marcel

Dekker, Inc.), ISBN 0-8247-9983-6.

Although great care has been taken to provide accurate and current information,

neither the author(s) nor the publisher, nor anyone else associated with this

publication, shall be liable for any loss, damage, or liability directly or indirectly

caused or alleged to be caused by this book. The material contained herein is not

intended to provide specific advice or recommendations for any specific situation.

Trademark notice: Product or corporate names may be trademarks or registered

trademarks and are used only for identification and explanation without intent to

infringe.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress.

ISBN: 0-8247-4696-1

This book is printed on acid-free paper.

Headquarters

Marcel Dekker, Inc., 270 Madison Avenue, New York, NY 10016, U.S.A.

tel: 212-696-9000; fax: 212-685-4540

Distribution and Customer Service

Marcel Dekker, Inc., Cimarron Road, Monticello, New York 12701, U.S.A.

tel: 800-228-1160; fax: 845-796-1772

Eastern Hemisphere Distribution

Marcel Dekker AG, Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland

tel: 41-61-260-6300; fax: 41-61-260-6333

World Wide Web

http://www.dekker.com

The publisher offers discounts on this book when ordered in bulk quantities. For

more information, write to Special Sales/Professional Marketing at the headquarters

address above.

Copyright � 2004 by Marcel Dekker, Inc. All Rights Reserved.

Neither this book nor any part may be reproduced or transmitted in any form or by

any means, electronic or mechanical, including photocopying, microfilming, and

recording, or by any information storage and retrieval system, without permission in

writing from the publisher.

Current printing (last digit):

10 9 8 7 6 5 4 3 2 1

PRINTED IN THE UNITED STATES OF AMERICA


www.dekker.com

FOOD SCIENCE AND TECHNOLOGY

A Series of Monographs, Textbooks, and Reference Books

EDITORIAL BOARD

Senior Editors Owen R. Fennema University of W isconsin-Madison Y. H. Hui Science Technology System Marcus Karel Rutgers University (emeritus) Pieter Walstra Wageningen University John R. Whitaker University of California-Davis

Additives P. Michael Davidson University of Tennessee-Knoxville Dairy science James L. Steele University of W isconsin-Madison Flavor chemistry and sensory analysis John H. Thorngate 111 University

food engineering Daryl B. Lund University of Wisconsin-Madison Food lipids and flavors David 6. Min Ohio State University Food proteins/food chemistry Rickey Y. Yada University of [Guelph Health and disease Seppo Salminen University of Turku, Finland Nutrition and nutraceuticals Mark Dreher Mead Johnson Nutritionals Phase transition/food microstructure Richard W. Hartel University of

Processing and preservation Gustavo V. Barbosa-Canovas Washington

Safety and toxicology Sanford Miller University of Texas-Austin

of California-Davis

W isconsin-Madison

State University-Pullman

1. Flavor Research: Principles and Techniques, R. Teranishi, 1. Hom- stein, P. Issenberg, and E. L. Wick

2. Principles of Enzymology for the Food Sciences, John R. Whitaker 3. Low-Temperature Preservation of Foods and Living Matter, Owen R.

Fennema, William D. Powrie, and Elmer H. Marth 4. Principles of Food Science

Part I: Food Chemistry, edited by Owen R. Fennema Part II: Physical Principles of Food Preservation, Marcus Karel, Owen R. Fennema, and Daryl B. Lund

5. Food Emulsions, edited by Sfig E. friberg 6. Nutritional and Safety Aspects of Food Processing, edited by Steven

R. Tannenbaum 7. Flavor Research: Recent Advances, edited by R. Teranishi, Robert A.

Flath, and Hiroshi Sugisa wa 8. Computer-Aided Techniques in Food Technology, edited by lsrael

SWJY


9. Handbook of Tropical Foods, edited by Harvey T. Chan 10. Antimicrobials in Foods, edited by Alfred Larry Branen and P. Michael

Davidson 11. Food Constituents and Food Residues: Their Chromatographic

Determination, edited by James F. Lawrence 12. Aspartame: Physiology and Biochemistry, edited by Lewis D. Stegink

and L. J. Filer, Jr. 13. Handbook of Vitamins: Nutritional, Biochemical, and Clinical Aspects,

edited by Lawrence J. Machlin 14. Starch Conversion Technology, edited by G. M. A. van Beynum and J.

A. Roels 15. Food Chemistry: Second Edition, Revised and Expanded, edited by

Owen R. Fennema 16. Sensory Evaluation of Food: Statistical Methods and Procedures, Mi-

chael O'Mahony 17. Alternative Sweeteners, edited by Lyn O'Brien Nabors and Robert C.

Gelardi 18. Citrus Fruits and Their Products: Analysis and Technology, S. V. Ting

and Russell L. Rouseff 19. Engineering Properties of Foods, edited by M. A. Rao and S. S. H.

Rizvi 20. Umami: A Basic Taste, edited by Yojiro Kawamura and Morley R.

Kare 21. Food Biotechnology, edited by Dietrich Knon 22. Food Texture: Instrumental and Sensory Measurement, edited by

Howard R. Moskowitz 23. Seafoods and Fish Oils in Human Health and Disease, John E.

Kinsella 24. Postharvest Physiology of Vegetables, edited by J. Weichmann 25. Handbook of Dietary Fiber: An Applied Approach, Mark L. Dreher 26. Food Toxicology, Parts A and B, Jose M. Concon 27. Modern Carbohydrate Chemistry, Roger W. Binkley 28. Trace Minerals in Foods, edited by Kenneth T. Smith 29. Protein Quality and the Effects of Processing, edited by R. Dixon

Phillips and John W. Finley 30. Adulteration of Fruit Juice Beverages, edited by Steven Nagy, John A.

Attaway, and Martha E. Rhodes 31. Foodborne Bacterial Pathogens, edited by Michael P. Doyle 32. Legumes: Chemistry, Technology, and Human Nutrition, edited by

Ruth H. Matthews 33. Industrialization of Indigenous Fermented Foods, edited by Keith H.

Stein kra us 34. International Food Regulation Handbook: Policy Science Law,

edited by Roger D. Middlekauff and Philippe Shubik 35. Food Additives, edited by A. Larry Branen, P. Michael Davidson, and

Seppo Salminen 36. Safety of Irradiated Foods, J. F. Diehl


37. Omega-3 Fatty Acids in Health and Disease, edited by Robert S. Lees and Marcus Karel

38. Food Emulsions: Second Edition, Revised and Expanded, edited by Ki re Larsson and Stig E. Friberg

39. Seafood: Effects of Technology on Nutrition, George M. Pigoff and Barbee W. Tucker

40. Handbook of Vitamins: Second Edition, Revised and Expanded, edited by Lawrence J. Machlin

41. Handbook of Cereal Science and Technology, Klaus J. Lorenz and Karel Kulp

42. Food Processing Operations and Scale-Up, Kenneth J. Valentas, Leon Levine, and J. Peter Clark

43. Fish Quality Control by Computer Vision, edited by L. F. Pau and R. Olafsson

44. Volatile Compounds in Foods and Beverages, edited by Henk Maarse 45. Instrumental Methods for Quality Assurance in Foods, edited by

Daniel Y. C. Fung and Richard F. Matthews 46. Listeria, Listeriosis, and Food Safety, Elliot T. Ryser and Elmer H.

Marth 47. Acesulfame-K, edited by D. G. Mayer and F. H. Kemper 48. Alternative Sweeteners: Second Edition, Revised and Expanded, ed-

ited by Lyn O'Brien Nabors and Robert C. Gelardi 49. Food Extrusion Science and Technology, edited by Jozef L. Kokini,

Chi-Tang Ho, and Mukund V. Kame 50. Surimi Technology, edited by Tyre C. Lanier and Chong M. Lee 51. Handbook of Food Engineering, edited by Dennis R. Heldman and

Daryl B. Lund 52. Food Analysis by HPLC, edited by Leo M. L. Nollet 53. Fatty Acids in Foods and Their Health Implications, edited by Ching

Kuang Chow 54. Clostridium botulinum: Ecology and Control in Foods, edited by

Andreas H. W. Hauschild and Karen L. Dodds 55. Cereals in Breadmaking: A Molecular Colloidal Approach,

Ann-Charlofte Eliasson and Ki re Larsson 56. Low-Calorie Foods Handbook, edited by Aaron M. Altschul 57. Antimicrobials in Foods: Second Edition, Revised and Expanded,

edited by P. Michael Davidson and Alfred Larry Branen 58. Lactic Acid Bacteria, edited by Seppo Salminen and Affe von Wright 59. Rice Science and Technology, edited by Wayne E. Marshall and

James 1. Wadsworth 60. Food Biosensor Analysis, edited by Gabriele Wagner and George G.

Guilbault 61. Principles of Enzymology for the Food Sciences: Second Edition, John

R. Whitaker 62. Carbohydrate Polyesters as Fat Substitutes, edited by Casimir C.

Akoh and Barry G. Swanson 63. Engineering Properties of Foods: Second Edition, Revised and

Expanded, edited by M. A. Rao and S. S. H. Rizvi


64. Handbook of Brewing, edited by William A. Hardwick 65. Analyzing Food for Nutrition Labeling and Hazardous Contaminants,

edited by Ike J. Jeon and William G. lkins 66. Ingredient Interactions: Effects on Food Quality, edited by Anilkumar

G. Gaonkar 67. Food Polysaccharides and Their Applications, edited by Alistair M.

'Stephen 68. Safety of Irradiated Foods: Second Edition, Revised and Expanded, J.

F. Diehl 69. Nutrition Labeling Handbook, edited by Ralph Shapiro 70. Handbook of Fruit Science and Technology: Production, Composition,

Storage, and Processing, edited by D. K, Salunkhe and S. S. Kadam 71. Food Antioxidants: Technological, Toxicological, and Health Perspec-

tives, edited by 0. L. Madhavi, S. S. Deshpande, and D. K. Salunkhe 72. Freezing Effects on Food Quality, edited by Lester E. Jeremiah 73. Handbook of Indigenous Fermented Foods: Second Edition, Revised

and Expanded, edited by Keith H. Steinkraus 74. Carbohydrates in Food, edited by Ann-Charlotte Eliasson 75. Baked Goods Freshness: Technology, Evaluation, and Inhibition of

Staling, edited by Ronald E, Hebeda and Henry F. Zobel 76. Food Chemistry: Third Edition, edited by Owen R. Fennema 77. Handbook of Food Analysis: Volumes 1 and 2, edited by Leo M. L.

Nollet 78, Computerized Control Systems in the Food Industry, edited by Gauri

S. Mittal 79. Techniques for Analyzing Food Aroma, edited by Ray Marsili 80. Food Proteins and Their Applications, edited by Srinivasan Damo-

daran and Alain Paraf 81. Food Emulsions: Third Edition, Revised and Expanded, edited by Stig

E. Friberg and K5re Larsson 82. Nonthermal Preservation of Foods, Gustavo V. Barbosa-Canovas,

Usha R. Pothakamury, Enrique Palou, and Barry G. Swanson 83. Milk and Dairy Product Technology, Edgar Spreer 84. Applied Dairy Microbiology, edited by Elmer H. Marth and James L.

Steele 85. Lactic Acid Bacteria: Microbiology and Functional Aspects: Second

Edition, Revised and Expanded, edited by Seppo Salminen and Atte von Wright

86. Handbook of Vegetable Science and Technology: Production, Composition, Storage, and Processing, edited by D. K. Salunkhe and S. S. Kadam

87. Polysaccharide Association Structures in Food, edited by Reginald H. Walter

88. Food Lipids: Chemistry, Nutrition, and Biotechnology, edited by Casimir C. Akoh and David B. Min

89. Spice Science and Technology, Kenji Hirasa and Mitsuo Takemasa


90. Dairy Technology: Principles of Milk Properties and Processes, P. Walsfra, T. J. Geurfs, A. Noomen, A. Jellema, and M. A. J. S. van Boekel

91. Coloring of Food, Drugs, and Cosmetics, Gisbert Offerstdffer 92. Listeria, Listeriosis, and Food Safety: Second Edition, Revised and

Expanded, edited by Elliot T. Ryser and Elmer H. Marth 93. Complex Carbohydrates in Foods, edited by Susan Sungsoo Cho,

Leon Prosky, and Mark Dreher 94. Handbook of Food Preservation, edited by M. Shafiur Rahman 95. International Food Safety Handbook: Science, International Regula-

tion, and Control, edited by Kees van der Heijden, Maged Younes, Lawrence Fishbein, and Sanford Miller

96. Fatty Acids in Foods and Their Health Implications: Second Edition, Revised and Expanded, edited by Ching Kuang Chow

97. Seafood Enzymes: Utilization and Influence on Postharvest Seafood Quality, edited by Norman F. Haard and Benjamin K. Simpson

98. Safe Handling of Foods, edited by Jeffrey M. Farber and €wen C. D. Todd

99. Handbook of Cereal Science and Technology: Second Edition, Re- vised and Expanded, edited by Karel Kulp and Joseph G. Ponte, Jr.

100. Food Analysis by HPLC: Second Edition, Revised and Expanded, edited by Leo M. L. Nollet

101. Surimi and Surimi Seafood, edited by Jae W. Park 102. Drug Residues in Foods: Pharmacology, Food Safety, and Analysis,

Nickos A. 8otsoglou and Dirnitrios J. Hetouris 103. Seafood and Freshwater Toxins: Pharmacology, Physiology, and

Detection, edited by Luis M. Botana 104. Handbook of Nutrition and Diet, Babasaheb B. Desai 105. Nondestructive Food Evaluation: Techniques to Analyze Properties

and Quality, edited by Sundaram Gunasekaran 106. Green Tea: Health Benefits and Applications, Yukihiko Hara 107. Food Processing Operations Modeling: Design and Analysis, edited

by Joseph lrudayaraj 108. Wine Microbiology: Science and Technology, Claudio Delfini and

Joseph V. Formica 109. Handbook of Microwave Technology for Food Applications, edited by

Ashim K. Daffa and Ramaswamy C. Anantheswaran 1 10. Applied Dairy Microbiology: Second Edition, Revised and Expanded,

edited by Elmer H. Marth and James L. Steele 11 1. Transport Properties of Foods, George D. Saravacos and Zacharias

B. Maroulis 1 12. Alternative Sweeteners: Third Edition, Revised and Expanded, edited

by Lyn O’Brien Nabors 113. Handbook of Dietary Fiber, edited by Susan Sungsoo Cho and Mark

L. Dreher 114. Control of Foodborne Microorganisms, edited by Vijay K. Juneja and

John N. Sofos I 15. Flavor, Fragrance, and Odor Analysis, edited by Ray Marsili


116. Food Additives: Second Edition, Revised and Expanded, edited by A. Larry Branen, P. Michael Davidson, Seppo Salminen, and John H. Thorngate, 111

1 17. Food Lipids: Chemistry, Nutrition, and Biotechnology: Second Edition, Revised and Expanded, edited by Casimir C. Akoh and David B. Min

118. Food Protein Analysis: Quantitative Effects on Processing, R. K. Owusu-Apen ten

119. Handbook of Food Toxicology, S. S. Deshpande 120. Food Plant Sanitation, edited by Y. H. Hui, Bernard L. Bruinsma, J.

Richard Gorham, Wai-Kit Nip, Phillip S. Tong, and Phil Ventresca 121. Physical Chemistry of Foods, Pieter Walstra 122. Handbook of Food Enzymology, edited by John R. Whitaker, Alphons

G. J. Voragen, and Dominic W. S. Wong 123. Postharvest Physiology and Pathology of Vegetables: Second Edition,

Revised and Expanded, edited by Jerry A. Bartz and Jeffrey K. Brecht 124. Characterization of Cereals and Flours: Properties, Analysis, and Ap-

plications, edited by Gonul KaletunG and Kenneth J. Breslauer 125. International Handbook of Foodborne Pathogens, edited by Marianne

0. Miliotis and Jeffrey W. Bier 126. Food Process Design, Zacharias B. Maroulis and George D. Sara-

vacos 127. Handbook of Dough Fermentations, edited by Karel Kulp and Klaus

Lorenz 128. Extraction Optimization in Food Engineering, edited by Constantina

Tzia and George Liadakis 129. Physical Principles of Food Preservation: Second Edition, Revised

and Expanded, Marcus Karel and Daryl B. Lund 130. Handbook of Vegetable Preservation and Processing, edited by Y. H.

Hui, Sue Ghazala, Dee M. Graham, K. D. Murrell, and Wai-Kit Nip 131. Handbook of Flavor Characterization: Sensory Analysis, Chemistry,

and Physiology, edited by Kathryn D. Deibler and Jeannine Delwiche 132. Food Emulsions: Fourth Edition, Revised and Expanded, edited by

Stig E. Friberg, K6re Larsson, and Johan Sjoblom

Additional Volumes in Preparation

Handbook of Frozen Foods, edited by Y. H. Hui, Paul Cornillon, Isabel Guerrero Legarrefa, Miang Lim, K. D. Murrell, and Wai-Kit Nip

Handbook of Food and Beverage Fermentation Technology, edited by Y. H. Hui, Lisbeth M. Goddik, Aase Solvejg Hansen, Jytfe Josephsen, Wai-Kit Nip, Peggy S. Stanfield, and Fidel Toldra

Industrialization of Indigenous Fermented Foods: Second Edition, Re- vised and Expanded, edited by Keith H. Steinkraus


Genetic Variation in Taste Sensitivity, edited by John Prescott and Beverly J. Tepper

Handbook of Food Analysis: Second Edition, Revised and Expanded: Volumes 1, 2, and 3, edited by Leo M. L. Nollet

Vitamin E: Food Chemistry, Composition, and Analysis, Ronald €itenmiller and Junsoo Lee


Preface to the Fourth Edition

Food Emulsions has now reached its fourth edition and very much reflectsthe strength of the original publication. Like the previous editions, this bookrealizes the value of the long tradition of diversity and excellence of researchand development within the food emulsion industry. This is exemplified byChapter 12 on beverage emulsions (Tan), by Chapter 2 on food emulsifiers(Krog and Sparsø); by Chapter 4 on proteins and polar lipids (Nylander); byChapter 1 on food emulsions in general (Dalgleish), and by Chapter 13 ondressings and sauces (Ford et al.).

There is probably no other emulsion category for which componentshave more influence on the properties of crystalline and liquid crystallinestructures, than lipids—Chapter 3 by Larsson on lipid structures is essentialreading.

The first edition of the book introduced advanced chapters on thefundamentals of food emulsion properties, and this aspect is a conspicuousfeature of the present book with Chapter 5 on destablilizing mechanisms,Chapter 6 on emulsion stablility, Chapter 9 on orthokinetic stability,Chapter 8 on coalescence mechanism, and Chapter 10 on the characteristicsof double emulsions.

Finally, this edition shows strength in an area not represented asstrongly earlier: namely, the different methods of characterization and ana-lysis of emulsions. Chapter 14 on droplet analysis by Coupland andMcClements, Chapter 7 on surface forces in emulsions by Claesson et al.,Chapter 11 on rheology of emulsions by Princen, and Chapter 15 on NMR infood emulsions by Balinov et al. give excellent overviews of these methods.


Needless to say, this book exceeds the quality of its predessors, andwe take this opportunity to recognize the truly outstanding efforts of ourcolleagues.

Stig E. FribergKare Larsson

Johan Sjoblom


Preface to the Third Edition

The economic and social changes during the last decades have changed theformulation requirements for emulsion systems in the most drastic manner.

Total cost analysis means that the selection of ingredients is no longerjust a question of cost per pound, but the efforts to stabilize the system mustnow be complemented by ‘‘hidden’’ costs for long-term technical or com-mercial failures—sometimes related only indirectly to stability. Social pres-sure has meant that new components with little or no nutritional value andwith intermolecular interactions different from traditional components mustbe accomodated, leading to phenomena for which the earlier methods pro-vide no appropriate response.

Taken in total, the consequences of change are that compled foodemulsion systems must be analyzed with proper attention to the colloidalstructures involved. Hence, the effects of the specific properties and inter-actions of polymers and proteins included in this book and the associationstructures of lipids leading to the formation of vesicles have received theattention they merit.

Stig E. FribergKare Larsson


Contents

Preface to the Fourth EditionPreface to the Third EditionContributors

1. Food Emulsions: Their Structures and Properties

Douglas G. Dalgleish

2. Food Emulsifiers: Their Chemical and Physical Properties

Niels J. Krog and Flemming Vang Sparsø

3. Molecular Organization in Lipids and Emulsions

Kare Larsson

4. Interactions Between Proteins and Polar Lipids

Tommy Nylander

5. Droplet Flocculation and Coalescence in Dilute Oil-in-Water

Emulsions

Øystein Sæther, Johan Sjoblom, and Stanislav S. Dukhin

6. Structure and Stability of Aerated Food Emulsions

D. T. Wasan, W. Xu, A. Dutta, and A. Nikolov

7. Surface Forces and Emulsion Stability

Per M. Claesson, Eva Blomberg, and Evgeni Poptoshev


8. Coalescence Mechanisms in Protein-Stabilized Emulsions

George A. van Aken

9. Orthokinetic Stability of Food Emulsions

Siva A. Vanapalli and John N. Coupland

10. Recent Developments in Double Emulsions for

Food Applications

Nissim Garti and Axel Benichou

11. Structure, Mechanics, and Rheology of Concentrated

Emulsions and Fluid Foams

H. M. Princen

12. Beverage Emulsions

Chee-Teck Tan

13. Dressings and Sauces

Larry D. Ford, Raju P. Borwankar, David Pechak,and Bill Schwimmer

14. Analysis of Droplet Characteristics Using Low-Intensity

Ultrasound

John N. Coupland and D. Julian McClements

15. NMR in Studies of Emulsions with Particular Emphasis

on Food Emulsions

Balin Balinov, Francois Mariette, and Olle Soderman


Contributors

Balin Balinov Department of Physical and Analytical Chemistry,Amersham Health, Oslo, Norway

Axel Benichou Casali Institute of Applied Chemistry, The HebrewUniversity of Jerusalem, Jerusalem, Israel

Eva Blomberg Royal Institute of Technology, Stockholm, Sweden

Raju P. Borwankar Kraft Foods, East Hanover, New Jersey, U.S.A.

Per M. Claesson Royal Institute of Technology, Stockholm, Sweden

John N. Coupland Department of Food Science, The Pennsylvania StateUniversity, University Park, Pennsylvania, U.S.A.

Douglas G. Dalgleish University of Guelph, Guelph, Ontario, Canada

Stanislav S. Dukhin New Jersey Institute of Technology, Newark, NewJersey, U.S.A.

A. Dutta Illinois Institute of Technology, Chicago, Illinois, U.S.A.

Larry D. Ford Kraft Foods, Memphis, Tennessee, U.S.A.

Nissim Garti Casali Institute of Applied Chemistry, The Hebrew Uni-versity of Jerusalem, Jerusalem, Israel

Niels J. Krog Danisco, Brabrand, Denmark

Kare Larsson Camurus Lipid Research, Lund, Sweden


Francois Mariette Cemagref, Rennes, France

D. Julian McClements The University of Massachusetts, Amherst,Massachusetts, U.S.A.

A. Nikolov Illinois Institute of Technology, Chicago, Illinois, U.S.A.

Tommy Nylander Department of Physical Chemistry, Center forChemistry and Chemical Engineering, Lund University, Lund, Sweden

David Pechak Kraft Foods, Glenview, Illinois, U.S.A.

Evgeni Poptoshev Royal Institute of Technology, Stockholm, Sweden

H. M. Princen Consultant, Flemington, New Jersey, U.S.A.

Øystein Sæther Department of Chemical Engineering, Norwegian Univer-sity of Science and Technology, Trondheim, Norway

Bill Schwimmer Kraft Foods, Glenview, Illinois, U.S.A.

Johan Sjoblom Department of Chemical Engineering, NorwegianUniversity of Science and Technology, Trondheim, Norway

Olle Soderman Department of Physical Chemistry, University of Lund,Lund, Sweden

Chee-Teck Tan Consultant, Middletown, New Jersey, U.S.A.

George A. van Aken Wageningen Centre for Food Sciences, Wageningen,The Netherlands

Siva A. Vanapalli Department of Chemical Engineering, University ofMichigan, Ann Arbor, Michigan, U.S.A.

Flemming Vang Sparsø Danisco, Brabrand, Denmark

D. T. Wasan Illinois Institute of Technology, Chicago, Illinois, U.S.A.

W. Xu Illinois Institute of Technology, Chicago, Illinois, U.S.A.


1Food Emulsions: Their Structuresand Properties

Douglas G. DalgleishUniversity of Guelph, Guelph, Ontario, Canada

I. INTRODUCTION

A. General Introduction

The taste and texture of a processed food perceived by the consumer dependon a variety of factors, important among which are the structures formed bythe constituent materials. The molecules which make up the food interact tocreate assemblies of molecules which give the structure and hence to a largeextent, determine the texture of the particular food. The ingredients areassembled during processing, and the structure created by the manufactureris governed by the controlled application of one or more effects: physical(e.g., interparticle forces, phase separations), chemical (e.g., formation ofspecific covalent bonds between molecules and particles), or biological (e.g.,fermentation, enzyme action). It is, of course, the aim of the processor togenerate products of predictable properties from materials whose propertiesare themselves understood and to do this as economically as possible.

Among the structures and structure-forming units within foods, emul-sions play a major part. They are known to impart desirable mouthfeelcharacteristics to the food, but, in addition, they are key ingredients inthe formation of structures in certain products, such as whipped toppingsand ice creams, and more complex products, such as processed cheeses.Therefore, the understanding of the formation, structures, and propertiesof emulsions is essential to the creation and stabilization of structures infoods. Food emulsions are widely used and are familiar to almost everyone.In addition to the products just mentioned, whole milk and cream are


emulsions, as are butter, margarine, spreads, mayonnaises and dressings,coffee creamers, cream liqueurs, some fruit drinks and whippable toppings.Many meat products depend on the presence of emulsions for theirproperties, as does bread dough, although in both cases the emulsionstructures can be extremely complex.

The formulation and creation of a food structure involving emulsionsis always a compromise, depending on the desired qualities of the food andthe materials which can be used to create these qualities. In addition to theessential physical functionality of the materials, it is necessary to take intoaccount nontechnological, but nonetheless important, factors. Foods con-tain ingredients which are subject to regulation by appropriate agencies. Insome cases, the use of certain ingredients may be discouraged because ofrestrictions imposed by certain religious groups or by public perceptions ofhealth issues. Furthermore, the processor is constrained by economics andcannot use expensive materials in product formulations. Last but not least,the product must be safe from a microbiological point of view; this mayhave important consequences because of the need for heat treatments, whichmay affect the stability of an emulsion during processing. All of these factorsmake the study of the efficient formulation and production of emulsions akey to the structure and behavior of processed foods.

B. Emulsion Types

An emulsion is a suspension of one phase in another in which it is immis-cible. One of the phases exists as discrete droplets suspended in the second,continuous, phase, and there is an interfacial layer between the two phaseswhich is occupied by some necessary surfactant material. There are threemain types of emulsion which are important, or potentially so, in foods. Inoil-in-water (o/w) emulsions, droplets of oil are suspended in an aqueouscontinuous phase. These are the most versatile of the emulsion types; theyexist in many forms (mayonnaises, cream liqueurs, creamers, whippabletoppings, ice cream mixes), and their properties can be controlled by varyingboth the surfactants used and the components present in the aqueous phase.Water-in-oil (w/o) emulsions are typified by butter, margarines, andfat-based spreads in general. These depend for their stability more on theproperties of the fat or oil and the surfactant used than in the properties ofthe aqueous phase, and because of this, there are fewer parameters whichcan be varied to control their stability. The third of the emulsion types iswater-in-oil-in-water (w/o/w), which is, in effect, an o/w emulsion whosedroplets themselves contain water droplets (i.e., are w/o emulsions). Theseare the most difficult emulsions to produce and control, because the water


droplets contained in the oil droplets must be stable, as must the oil dropletscontained in the continuous aqueous phase. These emulsions are describedin detail in Chapter 5 and will not be described further here.

For convenience of description, we may divide o/w food emulsionsinto three classes, depending on how they are to be used. The first classcontains emulsions which are end products in themselves. Coffee creamersand cream liqueurs are relatively simple emulsions whose only requirementis to remain stable toward creaming and coalescence during their shelf-life(which, however, may have to be considerable, so that sterility is also impor-tant). These emulsions present less of a challenge to the processor than domore complex emulsions; there are a few basic rules of formulation whichallow successful products to be created. The second class of emulsions con-tains those which can be used as ingredients that participate in forming thestructures of more complex products; that is, other components of the food(proteins, polysaccharides) form a matrix in which the fat globules aretrapped or with which they interact. Examples are yogurts, processedcheeses, and other gelled systems containing emulsion droplets whichmust interact with other components in the food, but are not destabilizedin the process. Their effect is to alter the rheological properties of the gel,thus creating texture and mouthfeel. In the third class of emulsion, thedroplets are required to create new structures during processing, such asin ice cream or whipped products (1,2), where the emulsion is destabilizedand further interacts as a means of creating structure in the product. Someemulsions themselves may also form gels during heating, to create newstructures within foods (3). The requirements for the compositions andproperties of the emulsion droplets are different in these different cases.However, it is generally necessary for the emulsion droplets to interactwith themselves and/or with the other food components to give the requiredstructures.

One of the ultimate goals in studying emulsions is to be able todescribe their functionality well enough from first principles to allow thereduction of the amount of fat in a product without, at the same time,adversely affecting its texture and organoleptic properties. Similarly, it isimportant to anticipate the possibility that emulsions will be used as carriersof flavors or bioactive materials. To best achieve this, we need to understandhow the fat functions in the original structure and how any material which isused to replace it should efficiently reproduce all of this function. Ideally, wewould like to be able to define the properties of a product and then toconstruct it from a knowledge of the possible ingredients. At the presenttime, this is possible for only a few of the simpler systems.

In the text which follows, emphasis has been given to emulsions pre-pared using milk constituents. This is simply because the milk proteins are


by far the most studied of the food proteins (being easily prepared, soluble,and relatively well behaved once they are in solution) and the emulsionsprepared using these materials have a correspondingly voluminous litera-ture, especially on the subject of ice cream. Reports on the basic propertiesof other proteins in emulsions are more fragmentary; this is not to say thatthey are unimportant, as the egg-based mayonnaise, to name only one, is avery widely used product.

II. THE BASIS OF THE BEHAVIOR OF OIL-IN-WATER EMULSIONS

A. General Aspects of Emulsions

Oil and water do not coexist comfortably because of the surface energy(Gibbs free energy) of the oil–water interface. Because of the interfacialtension between oil and water, any emulsion will seek to minimize theinterfacial energy by making the interfacial area between oil and water assmall as possible. In the absence of surfactants, this is achieved by coales-cence of the oil droplets, to give separated layers of oil and water. Thepresence of adsorbed surfactant molecules lowers the interfacial tensionbetween the oil and water phases, so that the driving force for coalescenceis reduced, although never to zero. Many surfactants (e.g., proteins) do notsimply reduce interfacial tension, but actively inhibit coalescence by alteringthe viscoelastic properties of the interface. The adsorbed material can alsoprevent the close approach of oil droplets by causing the surfaces to havesufficient charge to repel one another or by creating an extended surfacelayer, which also prohibits close approach. Thus, although emulsions tendto be regarded as thermodynamically unstable, it is possible, by judicioususe of surfactants, to control the kinetics of destabilization and to produceemulsions with very lengthy shelf lives. Surfactant molecules are amphiphi-lic; that is, they contain hydrophobic and hydrophilic domains. The formerdissolves in, or interacts with, the hydrophobic surface of the fat or oil,whereas the latter dissolves in the aqueous phase. The surfactant thereforeforms a layer on the surfaces of the emulsion droplets. Depending on thetype of surfactant, the adsorbed layer may have a complex structure,examples of which are described in the following subsections.

The thermodynamic instability leading to coalescence is, however,only one way in which emulsions can be unstable. Coalescence has themost drastic consequences, because it can be reversed only by rehomogeniz-ing the product, but other mechanisms which are important are creamingand flocculation. Both of these may promote coalescence and are generallynot to be favored. In creaming, the emulsion droplets do not lose theiridentity; they simply redistribute in space and can be returned to their


original state by agitation. Flocculation or aggregation arises from a morepermanent physical or chemical interaction between droplets. Flocs areoften not easily redispersed and may, therefore, have a negative effect onproduct quality (in soups, sauces, etc.). Because flocs are have a long life-time, the possibility that rupture of the interfacial layers and coalescence canoccur is enhanced.

The functional behavior of oil-in-water food emulsions is related totheir stability and is controlled by the three parts of the system: the fat or oilin the interior of the emulsion droplets, the interfacial material between thislipid material and the continuous aqueous phase, and the aqueous phaseitself. Each of these ‘‘phases’’ may be chemically complex. The lipid may bepartly or wholly crystalline and it may be subject to chemical change such aslipolysis or oxidation. The interfacial material can be composed of proteinsor of small emulsifiers such as monoglycerides, esters of fatty acids, orphospholipids, or mixtures of a number of these components. Finally, theaqueous phase may contain ions, which may interact with and potentiallydestabilize the emulsions, or macromolecules such as polysaccharides, whichmay exert either stabilizing or destabilizing effects. Therefore, to understandthe functional properties of the emulsions, it is necessary to understand theproperties of these three parts, individually and collectively.

B. Lipids and Emulsion Functionality

In oil-in-water emulsions, the fat or oil used to form the emulsion affects thefunctionality of the emulsion mainly by its degree of crystallinity, or itsability to crystallize. Oils which are liquid at the temperature at whichfoods are produced and consumed have little effect on the behavior of theemulsion, because they act essentially as fillers. They can, of course, coalesceif the fat globules are destabilized and the interfacial layer is sufficientlyweak, but they have little structural significance apart from that. On theother hand, it is possible for unsaturated liquid oils to undergo oxidation,and this, in turn, can lead to chemical reactions between the oxidized oil anda proteinaceous emulsifier (4). The overall effect of the reaction may be toalter functional properties of the emulsion and the nutritional value of thefood, as well as creating undesirable flavors. The details of how the func-tional properties of the emulsions are altered by these reactions are notknown, although it is known that the oxidized material is harder to displacefrom the interface than the original protein.

Emulsions are nearly always created (e.g., by homogenization) at atemperature at which all of the fat or oil is in a liquid state, and crystal-lization then occurs as the product is cooled to the temperature at which it isstored. Fats and oils which can crystallize in this way can be very important


in defining the functionality of the emulsion. By far the best known exampleof this is the involvement of partly crystalline fat in the mechanism of partialcoalescence, which stabilizes air bubbles in whipped emulsion products or inice creams (1). As the fat crystallizes, the growing crystals start to break theinterfacial layer of the fat globules (5), and this weakening of the surfacesallows more efficient destabilization of the emulsion (which in ice creammixes has often helped by weakening the interface by small-moleculesurfactants). In addition, the emulsion droplets, because their surfaces aredestabilized, are susceptible to being broken by the attachment to the air–water interfaces of the air bubbles as the mixture is whipped (6). The semi-liquid fat cannot wholly coalesce as the interfacial layers are broken,because of the limited mobility of the crystalline material, but it will partiallycoalesce to form a sintered layer around the bubble (7). As long as thecrystals are not melted, the layer will remain intact and stable around thebubbles. The fat may, of course, be completely crystallized after the partialcoalescence has occurred, but it must be semiliquid for the original phenom-enon to take place. Fats that are completely solid at the temperature atwhich whipping takes place are not efficient at stabilizing the foam.Equally, fat globule membranes that are too viscoelastic are too tough topermit the breakage essential for partial coalescence to occur.

Not all of the lipid components of a food are, however, neutral lipids inthe form of triglycerides. Phospholipids (lecithins) are a second class of lipidmaterials that are important in defining the properties of emulsions.However, in contrast to the fats and oils, which are present in the interiorsof the emulsion droplets, phospholipids are found in the interfacial layer (8)or may even not adsorb at all (9). It should be noted that although thelecithins are often described as ‘‘emulsifiers,’’ they are not as efficient ontheir own as are other small or large molecular emulsifiers. They may, never-theless, have a moderating effect on the properties of these other materials. Itis also probable that although it is popular to designate a whole range ofphospholipid materials under the heading of ‘‘lecithins,’’ each individualphospholipid type (with different head groups and fatty acids) will behavein a way unique to itself. Thus, phosphatidylcholines may behave differentlyfrom phosphatidylethanolamines, and distearyl phosphatidylcholine willbehave differently from dioleyl phosphatidylcholine. Hence, the source andcomposition of a lecithin sample will influence its functional properties.

C. The Interfacial Layer

The interfacial layers of many oil-in-water food emulsions contain proteins,which may be mixed with other surfactant materials (Fig. 1). Proteins areoften present in the raw materials of the food, and the fact that they are


excellent emulsifiers enhances their usefulness. The properties of the inter-facial layers depend not only on the quantities of materials adsorbed butalso on their properties, although we do not completely understand howthe composition and structure of an interfacial layer affect the detailedproperties of an emulsion.

The composition of the interfacial layer is governed mainly by what ispresent at the moment the emulsion is formed (10). If proteins are the onlyemulsifiers present, they will adsorb to the oil interfaces, generally in pro-portion to their concentrations in the aqueous phase (11). Certain mixturesof caseins are anomalous in this respect, because there is preferentialadsorption of b-casein from a mixture of purified as- and b-caseins (12).

Figure 1 Transmission electron micrographs of emulsion systems. (A) and (B)

show milk homogenized using a microfluidizer and centrifuged to separate different

populations of particles. (A) represents the larger fat globules, which float

during centrifugation (scale bar¼ 300 nm). (B) shows the sedimenting fraction

from the same milk, where very small fat globules are complexed with protein

particles (scale bar¼ 200 nm). (C) shows an emulsion made with soybean oil and

sodium caseinate, with only thin protein layers around the fat droplets (scale

bar¼ 200 nm).


However, this preferential adsorption does not seem to occur in emulsionsmade with sodium caseinate, where the caseins adsorb approximatelyaccording to their ratios in the original caseinate (11). Why this should beis not clear, although it may be a consequence of the different aggregatedstates of the caseins in solution (13). It is possible also that the concentrationof casein in the emulsion is important (14).

Small-molecule surfactants have important effects on the compositionof the interfacial layer. Depending on their nature, they may completelydisplace adsorbed protein, as in the case of water-soluble surfactantsadded after the formation of the interface (15), or partially displace theprotein, as found for oil-soluble surfactants, which must be added beforethe interface is formed (16). The effects of these surfactants may not beconfined to simply displacing the proteins; there is evidence for binding toproteins (17) or a complex displacement reaction, which has been observedby atomic force microscopy (18). Between the small-molecule surfactantsand proteins in size, there are peptides derived from the proteolytic break-down of protein molecules. These are also capable of stabilizing emulsions,although it appears that larger peptides are more effective at this thansmaller ones (19,20). Controlled proteolysis of proteins used as emulsifierscan give increases in their emulsifying efficiency; this has been observed forwhey proteins (9,21) and soy proteins (22,23).

The direct or competitive adsorption processes during the formationand storage of an emulsion are, of course, time and path dependent, asubject on which there is little information. This leads to difficulties wheninterpreting the properties of emulsions produced in laboratory conditions,where it is often the case that ingredients are mixed one at a time, with thenormal industrial situation, where many ingredients are mixed and pro-cessed at one time. Evidence of time dependence is manifest in the formationof networks of adsorbed whey proteins on the surfaces of emulsion dropletsas the emulsion is aged (24). The formation of disulfide linkages betweenadsorbed proteins is probably responsible for the stability of thatadsorbed layer, which is extremely difficult to displace (25). Strong rigidinterfacial layers can also be created by deliberate enzymatic cross-linkingof adsorbed proteins, the best known example of which is the use oftransglutaminase (26).

Details of the structures of the adsorbed layers will be discussed inSection VI. Briefly, a number of methods has been used to estimate thedimensions of adsorbed protein monolayers, among them are dynamiclight scattering (27), ellipsometry (28,29), and neutron reflectance (30).The results show that adsorbed layers of protein may be thick comparedto molecular dimensions. By forming a hydrodynamically thick layer andbecause they are generally charged, adsorbed proteins can stabilize emulsion


droplets by both steric repulsion and electrostatic (charge-repulsion)mechanisms.

Many, if not all, adsorbed proteins exist in conformations that aredifferent from their native states (31–33). This is a result of the tendencyof hydrophobic parts of the molecules to be adsorbed to the hydrophobicinterface, with a consequent distortion or disruption of their secondary ortertiary structures (34). As a result, the properties of the emulsion are notnecessarily the same as those of the parent protein. A more drastic manip-ulation of the adsorbed layer is possible—by the action of proteolyticenzymes on the adsorbed proteins. Although it would be difficult tocontrol on an industrial scale, the partial breakdown of adsorbed caseinby trypsin can considerably enhance the stability of the emulsion towardCa2þ (35).

D. The Continuous Phase

If the emulsion droplets are to contribute to the structure of a food, theymust interact in some way with the other components which are present.Interactions can occur between the droplets themselves, leading to gelationor flocculation, but other reactions are possible. If the components of thefood which are in the aqueous phase tend to form gels, then the emulsiondroplets may act in the simplest case as filler particles (i.e., they take upspace but do not interact physically or chemically with the gel) (36). On theother hand, the interfacial layer of the emulsion droplets may be capable ofinteracting with the aqueous-phase components as they gel (37); one obviousexample of this is the gelation of whey protein-based emulsions duringheating, where the protein in the aqueous phase interacts strongly withthe adsorbed whey protein of the emulsion droplet surfaces (38).Similarly, in the acid gelation of milk, which is part of the manufacture ofyogurt, the globules of milk fat are homogenized and end up with an inter-facial layer that is composed mainly of disrupted casein micelles (39). Thisallows the droplets to interact with free casein micelles as the acidificationproceeds. In yogurt, the interfacial layer remains intact after gelation, but inthe related product, cream cheese, the protein–fat emulsion gel is furtherworked, with the result that the interfacial layers are partially broken down,to give a different structure to the final product.

Interactions between emulsion droplets and macromolecules in solu-tion can be aided by the presence of certain ions, of which calcium is themost important. The presence of these ions may cause flocculation of theemulsions, or gel formation may be enhanced. A general increase in ionicstrength can destabilize the emulsions (40), but calcium may form morespecific bridges between emulsion droplets and materials in solution (41).


These effects can be greatly enhanced by inducing orthokinetic effects (i.e.,by stirring). In such a case, added ions may have a very strong effect (42).The basis of these effects is the fact that proteins generally carry a negativecharge in the pH region 5–7 and, therefore, are capable of binding the ions.In addition, the small ions alter the conformations and stabilizing propertiesof the adsorbed protein layer (43).

Polysaccharides in food systems containing emulsions behave in com-plex ways, which are being intensively researched at the present time. Theymay have at least three effects upon emulsions. First, there may be phase-separation effects, because of the thermodynamic incompatibility of thepolysaccharides, on the one hand, and the emulsion droplets, on the other(44). This leads to ‘‘depletion flocculation’’ processes, where the emulsiondroplets may be driven to form a concentrated emulsion phase, distinct fromthe aqueous phase containing the polysaccharide. Such depletion floccula-tion can also be caused by the presence of excess protein (45). This separa-tion can be more pronounced as the concentrations of emulsion andpolysaccharide are increased and may be rapid. Second, the polysaccharidemay gel, trapping the emulsion droplets and, in effect, stabilizing the emul-sion toward flocculation and coalescence (however, the presence of a suffi-cient quantity of polysaccharide to form a gel may also induce phaseseparation before gelation occurs). Third, the polysaccharides may interactdirectly with the adsorbed material on the surface of the emulsion droplet.Because polysaccharides are not hydrophobic in nature, they adsorb poorly,if at all, to lipid surfaces, although some galactomannans do show somesurface activity (46). In some cases, such as gum arabic, there is sufficientprotein associated with the gum to make it surface active (47–49); inaddition, synthetic protein–polysaccharide complexes have been shown tobe surface active (50,51).

In addition, charged polysaccharides may interact with adsorbed pro-tein of the opposite charge. It is possible to stabilize caseinate-based emul-sion droplets against acid precipitation by interaction with pectin (52),although the presence of pectin in emulsions can also lead to phase separa-tion (53). Caseinate emulsions can also be stabilized against acidificationby the presence of k-carrageenan, which may bind to the k-casein of thecaseinate even though both are negatively charged (52).

It is apparent that the possible interactions between an emulsiondroplet and the other components of a food can be very complex. Severalingredients are generally present and they may give rise to final structureswhich are dependent not only on overall composition, but also on themanner in which the ingredients are added and the time/temperaturevariations to which they are subjected during manufacture. Of the lasttwo, we have little knowledge, because the kinetics of exchange, phase


separation, and binding reactions have been studied very little, because ofthe complexity of the reactions and the products.

III. SURFACTANTS

A. Small-Molecule Surfactants

A considerable variety of surfactants is permitted for use in food emulsions,and these are discussed more fully in Chapter 8. Small-molecule surfactants(monoglycerides and diglycerides, sorbitan esters of fatty acids, polyox-yethylene sorbitan esters of fatty acids, phospholipids, and many others)contain long-chain fatty acid residues, which provide the hydrophobicgroup which binds to the lipid phase of the oil–water interface and causesadsorption. The head groups of these emulsifiers are more varied, rangingfrom glycerol (in monoglycerides and diglycerides) and substituted phos-phoglyceryl moieties (in phospholipids) to sorbitan highly substituted withpolyoxyethylene chains. Such material can have hydrophile–lipophilebalance (HLB) values from 3 (oil soluble) to 10 or higher (water soluble).As a general rule, emulsifiers of a low and high HLB are used to form w/oemulsions and o/w emulsions, respectively (54).

Because these molecules adsorb strongly to the oil–water interface andhave few steric constraints to prevent them from packing closely, they gen-erate low interfacial tensions (55) and are very effective at lowering theGibbs interfacial energy. However, they do not generally give highly cohe-sive or viscous surface layers, so that adsorbed layers of these small mole-cules may be quite easily disrupted (relative to adsorbed proteins; seeSection III.B). This property is indeed used in certain types of emulsion,where limited stability to coalescence is required.

B. Proteins

Proteins, on the other end of the scale of molecular complexity, act asemulsifiers but behave differently from the small molecules, because oftheir individual molecular structures, and, indeed, it is the particular pro-teins present which give many food emulsions their characteristic properties.Most, if not all, proteins in their native states possess specific three-dimen-sional structures (even though we may not know what they are) which aremaintained in solution, unless they are subjected to disruptive influencessuch as heating (56). When proteins adsorb to an oil–water interface, thehydrophobic regions of their structures (created by clusters of appropriateamino acid side chains) lie on, or possibly partially dissolve in, the oil phase.Some structures may be considered as especially important; for example, it is


possible for an a-helical portion of a protein to have a hydrophobic side,created by the hydrophobic side chains which lie outside the peptide core ofthe helix. However, even proteins such as caseins, which lack large amountsof regular structure, possess many groups of amino acids with hydrophobicside chains which adsorb to the oil–water interface. When a protein isadsorbed, the structure of the protein itself (the polypeptide backbone)will prevent close packing of the points of contact with the interface (theside chains), and as a result, adsorbed protein reduces the interfacial tensionless than do small molecules. Although some proteins are excellent emulsi-fiers, not all proteins can adsorb strongly to an o/w interface, either becausetheir side chains are strongly hydrophilic or because they possess rigidstructures that do not allow the protein to adapt to the interface [as inthe cases of gelatin (57) and of lysozyme, which although it does adsorbto o/w interfaces, tends to be a poor emulsifier (58)]. However, even appar-ently very hydrophilic proteins may adsorb strongly, as shown by the eggprotein phosvitin, which is a surprisingly good emulsifier (59–62) despitehaving more than 50% of its residues composed of phosphoserine (63), anamino acid which is charged and hydrophilic. In this case, even the relativelyfew hydrophobic residues in the protein are sufficient to cause adsorption,so that the protein can cover a large interfacial area with relatively fewpoints of contact.

Because adsorption of proteins occurs via the hydrophobic side chainsof amino acids, it has been suggested that a measurement of surface hydro-phobicity (64) should allow prediction of the emulsifying power of a protein(65). However, surface hydrophobicity is an ill-defined parameter, which isdetermined by the binding of probe molecules to the protein in solution (66)and may be a poor predictor of adsorption, especially because adsorbedproteins change conformation during or after adsorption.

C. Adsorption and Protein Conformation

Much research has been aimed at determining the mechanism of proteinadsorption, and it is likely that most of the proteins which adsorb well tointerfaces are capable of changing conformation either as they adsorb orshortly afterward. The concept of surface denaturation is well established(67,68) because the protein as it adsorbs is affected by spreading pressure,which pulls apart the native structure to maximize the amount ofhydrophobic contact with the oil interface (Fig. 2).

A number of methods have confirmed that proteins change their con-formations when they adsorb to liquid or solid interfaces. Spectroscopicstudies of lysozymes show that adsorption to polystyrene latex causes adecrease in the amount of secondary structure (69) and that the protein


may pass through a number of conformational states as the adsorptionprocess continues (70). The Fourier transform infrared (FTIR) spectra ofadsorbed a-lactalbumin and b-lactoglobulin have both been shown to differsignificantly from those of the native proteins (34,71). Proteins adsorbed toa surface and subsequently desorbed by the action of small molecules havebeen found to possess an altered conformation (72), showing that adsorp-tion-induced changes may be irreversible; for example, lysozyme and chy-mosin lose their enzymatic activity on adsorption and do not regain it afterbeing desorbed from the interface (32). It might be expected that the changein the conformation during adsorption is likely to destroy the secondary andtertiary structures of proteins, but it is possible to increase the orderedstructure in some cases (73).

Once adsorbed, some proteins are capable of interacting chemically byforming intermolecular disulfide bonds to give oligomers, as has been shownfor adsorbed b-lactoglobulin and a-lactalbumin (24,74), although such reac-tions do not occur in solution unless the proteins are denatured by heating(75). Further evidence for the denaturation of adsorbed proteins comes fromdifferential scanning calorimetry (DSC), suggesting that unfolding on thesurface has occurred (33,76,77). In some cases (e.g., lysozyme and a-lactal-bumin), this surface denaturation appears to be at least partially reversible,but in others (e.g., b-lactoglobulin), adsorption causes irreversible changesin the protein molecules.

The casein proteins tend to be a special case. Because these proteinsappear not to contain much rigid secondary structure (a-helix or b-pleated

1 2

3 4

5 6

Oil

Water

Figure 2 Schematic diagram of the adsorption and desorption of protein. The

surface of the protein has hydrophobic (dark) and hydrophilic (light) regions. The

protein molecule approaches the interface (1) and begins to adsorb. In principle, it is

possible that very rapid desorption may take place (2) without the protein changing

conformation. With time, the adsorbed protein changes its conformation to

maximize hydrophobic contact with the oil, and this may pass through several

stages (4). At this stage, the protein may be hard to displace (dotted arrows), and

even if it is displaced, it will have an altered conformation (5) or even be denatured

and aggregated (6). The adsorbed protein may itself react with its neighbors to form

a network.


sheet) (78) and because they possess considerable numbers of hydrophobicresidues (79), they adsorb well (80). However, because of the lack of defini-tion of their original native structures, it is impossible to determine whetherconformational changes occur during adsorption, as neither spectroscopicchanges nor DSC are capable of demonstrating conformational changes inthese proteins.

Reactions between adsorbed protein molecules in emulsions (forinstance, disulfide bridging interactions such as those mentioned earlier)will be encouraged by the very high local concentration of protein withinthe adsorbed interfacial layers. Generally, we know (81,82) that for mono-layers of adsorbed proteins, the interfacial concentration (surface excess, �)is generally between 1 and 3mg/m2 and that the adsorbed layers are gen-erally less than 5 nm thick (83), so it is simple to calculate that, in theinterfacial region, a monolayer of protein has an effective concentrationof about 500mg/mL (i.e., 50%). This is, of course, very much higher thancan be achieved by attempting to directly dissolve the proteins because ofthe extremely high viscosity generated, so direct comparisons betweenadsorbed and unadsorbed proteins at equal effective concentrations arenot possible. However, the protein in the adsorbed layer may be in a favor-able position for intermolecular interactions, because the molecules are veryclose to one another and adsorption holds them in position so that diffusionis slow. We should probably regard the adsorbed layer of protein as beingmore like a gel than a solution; this is at least partly the reason why manyadsorbed proteins form highly viscous interfacial layers. These gels will beessentially two dimensional, with each molecule occupying approximately11 nm2 of interface [calculated on the basis of a molecule of 20,000Da and asurface coverage of 3mg/m2; this agrees well with the expected dimensionsof a globular protein of this weight (84) and is much larger than the0.5–2.5 nm2 per molecule which has been found for adsorbed modifiedmonoglycerides (85)]. It is, therefore, not surprising that adsorption canalter the structures of proteins. The formation of such concentratedlayers has relatively little to do with the overall bulk concentration of theprotein in solution, which may give stable emulsions at relatively lowbulk concentrations (although this depends on the amount of oil and theinterfacial area to be covered).

Caseins form extended layers about 10 nm thick, and even at a � of3mg/m2 have a ‘‘concentration’’ of about 300mg/mL. Conversely, wheyproteins form much thinner layers (about 2 nm thick) and must begin toform multilayers if � is more than about 2mg/m2, as there is no furtherspace available for monolayer adsorption beyond that point.

With a few exceptions, most of the detailed research has been per-formed on relatively few proteins. Of these, the caseins (as1, as2, b, and k)


and whey proteins (a-lactalbumin and b-lactoglobulin) predominate. This isprincipally because these proteins are readily available in pure and mixedforms in relatively large amounts; they are all quite strongly surfactant andare already widely used in the food industry, in the form of caseinates andwhey protein concentrates or isolates. Other emulsifying proteins are lessamenable to detailed study by being less readily available in pure form (e.g.,the proteins and lipoproteins of egg yolk). Many other available proteins areless surface active than the milk proteins [e.g., soya isolates (86)], possiblybecause they exist as disulfide-linked oligomeric units rather than as indivi-dual molecules (87). Even more complexity is encountered in the phosphory-lated lipoproteins of egg yolk, which exist in the form of granules (88),which themselves can be the surface-active units (e.g., in mayonnaise) (89).

IV. FORMATION OF EMULSIONS AND MEASUREMENT OFEMULSIFYING ACTIVITY OF PROTEINS

Food oil-in-water emulsions are generally produced using either colloidmills or high-pressure homogenizers. In the former, the oil–water–surfactantmixture is passed through a narrow gap between a rotor and stator, in whichthe stresses imposed on the mixture are sufficient to break up the oil intodroplets, to which the surfactant adsorbs. This method tends to producedroplets of emulsion which are larger than those produced by high-pressurehomogenization, being of the order of 2 mm in diameter. The technique isused to manufacture mayonnaises and salad creams, in which stabilitydepends less on the presence of very small particles than on the overallcomposition and high viscosity of the preparation. In liquid emulsions,however, smaller particles are required to prevent creaming and possiblecoalescence.

High-pressure homogenization is used to produce these smallerdroplets. A coarse emulsion of the ingredients is formed by blending, andthis suspension is then passed through a homogenizing valve, at pressureswhich are generally in the region of 6.8–34MPa (1000–5000 psi). This high-pressure flow through the valve creates turbulence, which pulls apart the oildroplets, during and after which the surfactant molecules adsorb to thenewly created interface (90). If the adsorption is not rapid, or if there isinsufficient surfactant present to cover the freshly formed interface, thenrecoalescence of the oil droplets rapidly occurs (91). The breakup and recoa-lescence occurs many times as the droplets pass through the field generatedby the homogenizer (92). Apart from the mechanical design of the homo-genizer, the sizes of the emerging droplets depend on, among other factors,the homogenization pressure, the viscosity of the suspension, the number of


passes (93), and the amount and types of surfactant present (94). Generally,when the surfactant is present in excess concentration, the particle size islimited by the characteristics of the homogenizer and of the suspension; onthe other hand, if only small amounts of surfactant are present, the surfac-tant concentration limits the sizes of the particles, because insufficientlycovered emulsion droplets will recoalesce. Generally, therefore, as thecompositions of products are reformulated, the sizes of the emulsiondroplets in them will change.

In addition to recoalescence and increased droplet size in the pre-sence of insufficient surfactant, the phenomenon of bridging flocculation,in which the emulsion droplets form clusters during homogenization, canbe observed. For the bridging to occur, it is necessary to have macromo-lecular surfactants with at least two sites by which they can adsorb tointerfaces. At low surfactant concentrations, such molecules can becomeadsorbed to two separate oil droplets. Proteins can form bridges in thisway (95), and even more commonly, natural aggregates of proteins such ascasein micelles can induce clustering of the oil droplets (96). Bridgingflocculation may be reversed by incorporating more surfactant (whichneed not be macromolecular) so as to provide enough material to coverthe interface as it is formed. In the case of clustering by particles whichthemselves can be broken up, a second-stage homogenization at lowerpressure can be sufficient to break down the bridging aggregates and toseparate the clustered fat globules. Clearly, however, such treatment willbe inapplicable to clusters bridged by single macromolecules, which cannotbe broken up in this way.

One factor which can have considerable importance on the emulsifyingproperties of proteins is their quaternary structure. For example, in milk,the caseins exist in aggregates of considerable size (casein micelles) contain-ing hundreds or thousands of individual protein molecules (97), heldtogether by hydrophobic interactions and microparticles of calcium phos-phate. The casein micelles act as the surfactants when milk is homogenized(98). During this homogenization, the micellar structure is disrupted, pos-sibly by the forces within the homogenizer (99), but presumably also by thespreading forces which occur when the micelles violently encounter the oilsurface. The result is that the oil surface is unevenly coated by partiallybroken up casein micelles, and not by a monolayer of casein (Figs. 1Aand 1B).

In contrast, sodium caseinate (which is prepared by removing thecalcium phosphate from the micelles by precipitation at acid pH and thenwashing the precipitate and redissolving at neutral pH) has much superioremulsifying properties compared to casein micelles (100) (i.e., the amount ofoil which can be stabilized by a given weight of casein in either of the two


forms, under identical homogenization conditions). This effect is probablysimply because the effective concentration of emulsifier is much less whenthe casein is in the micellar form, which is relatively resistant to disruption.Therefore, during homogenization, the nonmicellar casein will arrive at theinterface more readily than the micelles. Interestingly, sodium caseinate atthe concentrations generally used (above about 0.5% w/w protein) is notitself monomeric, but exists in the form of aggregates of the proteins con-taining about 30 molecules (13), which are held together probably by hydro-phobic forces. In contrast to the intact micelles, these particles are believedto create monolayers of casein molecules around the fat globules; that is, theaggregates are pulled apart by the spreading pressure which they encounteras they bind to the interface. This cannot happen to casein micellar frag-ments, whose integrity is probably maintained by the presence of calciumphosphate.

Molecules such as b-lactoglobulin also show changes in quaternarystructure as a function of pH (101), and these may be related to the changesin the protein’s surfactant properties at different pH values (102). The dena-turation of b-lactoglobulin by heat causes the protein to aggregate, and thisdecreases the emulsifying power to a considerable extent (103).

Because different proteins are more or less efficient at forming andstabilizing emulsions, and even the same protein may have different efficien-cies in different circumstances (as has just been described for casein), it isessential to have methods for estimating the potential of given surfactantsfor forming emulsions. To achieve this, the required techniques should bemethod independent; that is, they should give absolute results, or at leastgive results applicable to specific methods for preparing emulsions. Thereare two widely used methods, Emulsifying Activity Index (EAI) andEmulsifying Capacity (EC). Neither of these methods is method indepen-dent, although they are simple to apply. To measure EC, a known quantityof surfactant is dissolved in water or buffer and then oil is added to it in ablender. This forms a crude emulsion, and further aliquots of oil are addeduntil the emulsion inverts or free oil is seen to remain in the mixture. Thisostensibly gives the weight of oil, which can be emulsified by the definedweight of protein. It is evident that this method is dependent on the parti-cular blender because what is important in emulsion formation is not theweight of oil per se but its interfacial area. Thus, if the emulsion is made oflarge droplets, it will consume less surfactant than if small droplets arepresent. The conditions of emulsion formation are therefore critical to themethod, as it is possible to obtain different results at different blenderspeeds, or with other homogenizing devices. Therefore, the method is notin any sense an absolute measure. As a quality control measure or as aninternal method in a single laboratory, it may have considerable usefulness.


To measure the EAI, an emulsion is made and the particle size of theemulsion droplets is estimated. The assumption is then made that all of theprotein is adsorbed to the interface, and so a measure of emulsifying poten-tial can be measured. Although it provides more information than EC, themethod has two major defects: First, it is very often the case that not all ofthe available protein is adsorbed, or adsorbed as a monolayer. Indeed, it isknown that at concentrations of protein of more than about 0.5% (with oilconcentration of 20%), some of the protein remains unadsorbed, even afterpowerful homogenization where the concentration of protein is the limitingfactor in the determination of the sizes of the droplets (11,94). If homoge-nization is less extensive, then the proportion of protein which is adsorbeddecreases. The second major problem in interpretation of the EAI is simplythe difficulty of determining the particle sizes and their distribution. Thereare a variety of methods for measuring the size distribution of suspendedparticles, and care must be taken to avoid error in this measurement.Traditionally, the particle sizes in determinations of the EAI are measuredby determining the turbidity of diluted suspensions of the emulsions, whichis a method much subject to error.

Ideally, to fully describe the emulsifying capacity of a surfactant, theparticle size distribution of the emulsion and the amounts of individualsurfactants adsorbed to the oil–water interface need to be measured. It ispossible to measure the amount of adsorbed protein by centrifuging theemulsion so that all of the fat globules form a layer above the aqueousphase and measuring by chromatography the concentration of surfactantleft in the latter phase (12). Alternatively, the fat layer after centrifugationcan itself be sampled, and the adsorbed protein can be desorbed from theinterface by the addition of sodium dodecyl sulfate (SDS) and quantified byelectrophoresis on polyacrylamide gels (104). In addition, although this ismore difficult to determine, it is desirable to know the state of the adsorbedmaterial (e.g., its conformation, which parts of the adsorbed moleculesprotrude into solution and are available for reaction, etc.). This representsan ideal which is rarely possible to achieve, but the explanation of thebehavior of emulsions and perhaps the design of new ones may dependon this knowledge.

V. MEASUREMENT OF PARTICLE SIZES AND SIZEDISTRIBUTIONS IN EMULSIONS

Once an emulsion has been formed using homogenization or other means,it is often necessary to characterize it, specifically in terms of itsaverage size and its size distribution. This is important in a number of


respects: Knowledge of the size distribution provides information on theefficiency of the emulsification process, and the monitoring of any changesin the size distribution as the emulsion ages gives information on the stabi-lity of the system. Regular measurement of particle size can be part of aquality control operation and can also be important when emulsion systemsor processes are patented. However, the measurement of true size distribu-tions or even the average sizes of emulsion droplets is not simple, despite theexistence of a number of potentially useful and apparently simple methods.

The most direct method and one which is theoretically least subject toerrors is electron microscopy (105). This technique measures the number-average size distribution, providing that (a) a fully representative sample ofthe emulsion is prepared, fixed, mounted, sectioned, and stained withoutdistortion, (b) a sufficient number of particles is measured to ensure statis-tical accuracy of the distribution, and (c) proper account is taken of theeffects of sectioning on the apparent size distribution. All of this requiresconsiderable time, effort, and calculation, so that the technique cannot beused routinely to determine size distributions. It may be used as a standardagainst which to compare other methods, and it also finds a use in measur-ing systems where dilution causes changes in the particle sizes, as in micro-emulsions (106).

The most widely used of the rapid methods for particle sizing arebased on light scattering. These tend to emphasize large particles in thedistribution because larger particles scatter more light than smaller ones.The simplest of these methods depends on the measurement of turbidity atone or a number of wavelengths (107,108). Turbidity, or apparent absor-bance of light, is a measure of the total amount of light scattered as it passesthrough a cuvette containing diluted emulsion (assuming that no componentof the emulsion absorbs light of the wavelengths used). Although themethod is rapid and may be performed in any laboratory possessing aspectrophotometer, it cannot be used to give the true distribution of particlesizes, but at best to give an average. It can be assumed that the particlesform a distribution of known shape, but this, of course, assumes that thedistribution is known beforehand.

A number of commercial instruments measure the distributions ofparticle sizes by determining the intensity of light scattered from a highlydiluted sample at a number of specific scattering angles [integrated lightscattering (ILS)]. With knowledge of the scattering properties (i.e., theMie scattering envelope) of the particles (109), software is used to calculatethe most probable distribution of particle sizes.

This does not always yield the true absolute distribution, for two mainreasons. The first of these is that the angular range is often too restricted toallow measurement of small particles of diameters less than about 50 nm,


which scatter almost isotropically. Large particles preferentially scatter in theforward directions, so that to measure the distribution accurately, ideallymeasurements have to be made at a span of scattering angles between 0� and150� (110). This is now available in modern particle sizers, but in olderinstruments, the angular range used limits the detection of small particles(smaller than 0.1 mm). Many food emulsions made by high-pressure homo-genization contain particles of this size within their size distribution. Afurther problem with any light-scattering method is that the accuracy ofthe calculated distribution depends on how well the optical properties ofthe emulsion droplets (i.e., their real and imaginary refractive indices, whichdetermine the scattering properties) can be defined. Generally, it is realisticto assume that the emulsion droplets are spherical, but it may be necessaryto make assumptions about the structures of the interfacial layers. An emul-sion droplet is essentially a coated sphere (111), which is characterized byrefractive indices of the core and the coat, and these are likely to differ.Calculations based on of the scattering behavior of emulsion dropletsmay, therefore, depend on the presumed structures and compositions of theparticles.

If the emulsion is unstable, the particle size distribution will changewith time, and this will be detected by the light-scattering measurements.However, a simple measure of light scattering cannot distinguish betweendroplets which have flocculated and those which have coalesced, and othermethods of measurement are needed to define which type of instability hasoccurred. Flocculation introduces another problem relative to the detailedinterpretation of light-scattering measurements, because it produces parti-cles which are neither homogeneous nor spherical. To determine the type ofinstability which has occurred, it is often possible to use light microscopy.Alternatively, the destabilized emulsion can be treated with SDS, which willdissociate any flocs. A second measurement of the particle size after the SDStreatment will show no change if the emulsion had coalesced, but it willrevert to the original particle size distribution if the destabilization has beenby flocculation (112).

Dynamic light scattering (DLS) offers an alternative means of mea-surement (113). This technique does not measure the total amount of lightscattered, but the dynamics of the scattered light over very short time peri-ods. Usually, the light scattering is measured at a fixed angle of 90� and acorrelation function is measured. This is essentially a weighted sum of expo-nentials, which depend on the diffusion coefficients of the particles throughthe aqueous medium. As with ILS, the calculation of the true size distribu-tion depends on the knowledge of the detailed light-scattering properties ofthe emulsion droplets. In addition, the fit of theory to the true correlationfunction is ill-conditioned (114), so that the size distribution obtained can


depend on the technique used to fit the correlation function. As with otherlight-scattering techniques, the contribution of larger particles to the sizedistribution is generally overestimated. This is partly because they tend toscatter more (i.e., have higher weighting factors), but also because of thenature of the correlation function itself, as the information about the smallparticles is contained only in the short-time part of the function, whereasinformation about the large particles is contained at all points.

An important demand of both ILS and DLS is that they require thesuspensions of particles to be highly diluted. This is necessary because thetheories used to calculate the particle sizes demand that the scatteredphotons have undergone only one scattering event. Multiple scattering dis-torts the photon statistics and leads to erroneous results. The dilution is notnecessarily serious in the case of simple stable emulsions, but in more com-plex emulsions, it is necessary to ensure that no dissociation of particles iscaused by the high dilution required for accurate light-scattering experi-ments. The dilution may lead to dissociation of flocculated material or tothe breakdown of complex interfacial layers (e.g., those formed by caseinmicelles on the oil–water interfaces in homogenized milk). Conversely, it ispossible that dilution into an inappropriate solution may promote aggrega-tion or the emulsion droplets (e.g., dispersing casein-stabilized droplets in asolution containing large concentrations of calcium ions).

Recent advances in instrumentation have allowed some measurementsof DLS to be made on more concentrated suspensions. Three methods havebeen developed. In the first, the optical paths of the incident and scatteredbeams of the instrument are positioned so that the measurements take placejust inside the wall of the cuvette containing the sample. In such a case, thepath lengths of the incident and scattered photons in the sample are veryshort, and up to a volume concentration of as much as 20%, the system maybe considered as a simple DLS experiment with no multiple scattering. Asecond, more complex, but in principle more accurate, method is to use twoscattering beams which are cross-correlated with one another (115). Thiseffectively compares two scattering experiments on the same sample andallows the elimination of the contribution of multiply-scattered photons.This technique has been used to study milk (116) but not, so far, emulsions.Finally, a third technique uses the multiply-scattered light rather than tryingto eliminate its effects. This is the technique of diffusing wave spectroscopy(DWS). The method is useful for concentrated suspensions such as emul-sions (117), but although it can give average sizes of the particles, distribu-tions are harder to obtain. It has, however, been more widely used than theother methods, especially in empirical ways, to detect gelation (118).

These three methods are not as yet widely used, because instrumenta-tion is only now coming on to the market, and most DWS apparatuses tend


to be custom-built in a single laboratory. In using them, it is essential also tobe aware of a problem that is not present in diluted solutions. All of themethods of dynamic light scattering depend on the presence of movingparticles; in fact, they measure diffusion coefficients, which are then trans-formed into particle sizes. In turn, diffusion coefficients depend on theviscosity of the continuous phase. When concentrated suspensions aremeasured, the particles themselves contribute to the viscosity of the sus-pension, so that care must be taken when evaluating particle sizes frommeasurements made in concentrated suspensions.

Light-scattering methods are, at present, the most effective and mostwidely used means of obtaining information about the size distributions ofparticles in emulsion systems. Like many methods, they are especially usefulin a comparative mode to measure changes that occur during processing orstorage of the suspensions. All of the light-scattering methods can detectwhether aggregation is occurring, so that they may all be used to detectinstability of emulsions, and they are almost unique in allowing the kineticsof aggregation to be studied on a real-time basis (119). Simply, the fact thatthe particle size is increasing can be determined without any particularattributes of the particles needing to be known.

An alternative method of particle sizing, which shows considerablepotential although it is not so widely used as light scattering, is based onultrasonic acoustic spectroscopy. This measurement can easily be made onconcentrated dispersions and depends on the fact that the attenuation andvelocity of ultrasound of a defined frequency through a suspension dependson the sizes of the particles in the dispersed phase. By measuring theattenuation of sound at a series of different frequencies through thesample, it is possible to calculate the size distribution of the particles inthe suspension (120,121). Instruments to perform the measurements areavailable and they are capable of being compared with the informationavailable from light-scattering measurements (122). Just as the scatteringof light depends on the relative refractive indices of the dispersed phaseand the continuous phase, so particle sizing by ultrasound also dependson the physical properties of the dispersed phase; for example, the density,viscosity, thermal conductivity, and specific heat are all required to beknown to permit the true size distribution to be determined (123). In theabsence of these factors, ultrasonic attenuation spectroscopy can give onlyrelative size distributions. However, because of the applicability of themethod to real (undiluted) emulsion systems, it is likely that the methodwill increase in its usage. Ultrasound and its applications are discussed indetail in Chapter 10.

A related method which has been proposed is that of electroacousticspectroscopy (124). Because the passage of ultrasound through a dispersion


disturbs the double layer which surrounds the dispersed particles, an electriccurrent is set up. Measurement of this current allows the calculation also ofthe �-potential of the particles as well as their size distribution. The mea-surement may also be performed in reverse; an oscillating electric fieldcauses the emission of ultrasound, which, in turn, permits the size distribu-tion and the �-potential to be measured (125,126). Like the simpler ultra-sonic spectroscopy, these methods require precise knowledge of the physicalcharacteristics of the medium and the dispersed phase if they are to giveabsolute, rather than relative, results.

VI. STRUCTURES OF THE ADSORBED LAYERS ON THESURFACES OF EMULSION DROPLETS

A. Simple Surfactants

The structures of the interfacial layers in emulsion droplets might beexpected to be simple when small-molecule emulsifiers are used, but this isnot necessarily the case, especially when not one but a mixture of surfactantmolecules is present. Although simple interfacial layers may be formedwhere the hydrophobic moieties of the surfactants are dissolved in the oilphase and the hydrophilic head groups are dissolved in the aqueous phase, itis also possible for multilayers and liquid crystals or even crystals to formclose to the interface. This depends on the nature and the concentrations ofthe different surfactants. Interactions between surfactants generally enhancethe stability of the emulsion droplets, because more rigid and structuredlayers tend to inhibit coalescence. Also, mixtures of different surfactantshaving different HLB numbers appear to provide structured interfaciallayers, presumably because of the different affinities of the surfactants forthe oil–water interface. Specifically, phospholipids may form multilamellarstructures around the oil–water interface, and presumably these layers willhave different spacing depending on the amount of hydration (127). Theyalso depend both on the nature of the oil and of the phospholipid (128). So,although the major adsorption of phospholipids at low concentration islikely to be in the form of monolayers, it may be possible to producemore complex structures when large amounts of phospholipid are presentor when other surfactants are coadsorbed to the interface.

B. Interfaces with Adsorbed Proteins

Undoubtedly, the most complex interfacial structures are produced whenproteins are used as the surfactants, because of the great range of conforma-tional states accessible to such molecules. This is of interest because of the


implications of conformational change on the reactivity and functionality ofthe proteins [e.g., it appears that adsorbed b-lactoglobulin cannot formdisulfide bonds with k-casein when heated in the presence of caseinate(129), although this reaction is known to occur between the proteins insolution and in heated milk]. Flexible molecules such as caseins may beconsidered as adsorbing as if they were heteropolymers (130) because theyare presumed to have high conformational mobility (78). As evidence of thepresumed conformational change, adsorbed b-casein exhibits differentsusceptibility to attack by proteolytic enzymes, compared with the proteinin solution (131). Further, adsorption to different hydrophobic materialscauses differences in the conformation of the adsorbed molecule; forexample, the protein seems to have somewhat different conformationswhen adsorbed to hydrocarbon (n-tetradecane) or triglyceride (soya oil)–water interfaces (132). As a result of these measurements, it appearsthat model hydrocarbon–water systems are not necessarily suitable fordescribing triglyceride–water systems.

Much is known about the structure of adsorbed b-casein, certainlymore than is known for any other food protein. The first evidence fromdynamic light scattering showed that b-casein can adsorb to a polystyrenelatex and cause an increase in the hydrodynamic radius of the particle by10–15 nm (133). Small-angle x-ray scattering confirmed this and showedthat the interfacial layer was not of even density throughout and that thebulk of the mass of the protein was close to the interface (134). Neutronreflectance studies also showed that most of the mass of the protein wasclose to the interface (135). From these results, we can infer that a relativelysmall portion of the adsorbed protein molecule extends from the tightlypacked interface into solution, but it is this part which determines thehydrodynamics of the particle and which must be the source of the stericstabilization which the b-casein affords to emulsion droplets (133). All of thestudies just described were performed on polystyrene latex particles or onplanar interfaces; however, it has also been demonstrated that the interfacialstructures of b-casein adsorbed to emulsion droplets resemble those in themodel particles (83,134). Although detailed control of emulsion dropletsduring their formation in a homogenizer is impossible, it is possible tobreak down the surface layers of protein once they are formed, by the useof proteolytic enzymes (135), and by comparison with the behavior of modelsystems under similar conditions, it is possible to demonstrate that theproteins seem to have similar conformations in the model systems andemulsions (27).

It is also possible to use proteolytic enzymes to demonstrate whichpart of the b-casein protrudes into the solution. There are many sites inthe protein molecule (lysine and arginine residues) susceptible to attack by


trypsin, and these are almost equally susceptible to attack when the proteinis free in solution. In the adsorbed protein, sites close to the N-terminal aremost readily accessible, presumably because they form the part of theadsorbed layer which protrudes into solution (131). This region of the mole-cule is the one which would be expected to behave in this way, being themost hydrophilic and highly charged part of the protein. Thus, for b-casein,it is possible to predict some features of the conformation of the adsorbedprotein from a study of its sequence (131). It seems that b-casein is perhapsthe only protein for which this kind of prediction can be done; most otherproteins (even the other caseins) have much less distinct hydrophilic andhydrophobic regions and, therefore, have conformations which are moredifficult to predict (27). From studies of the structure of adsorbed as1-caseinn model systems and in emulsions, it is established that neither the mostaccessible sites for trypsinolysis (136) nor the extent of protrusion of theadsorbed protein into solution (137) can be readily predicted.

Nevertheless, it is possible to calculate from statistical mechanicalprinciples the approximate conformations of the adsorbed caseins, byassuming that they are flexible, composed of chains of hydrophilic andhydrophobic amino acids (138). The results of these calculations reproducemany of the features of the actual measured properties, especially the ten-dency of the adsorbed b-casein to protrude further from the interface thanthe as1-casein (139). These calculations have, in turn, been used to explainthe differing stabilities of the two different types of emulsions toward addedsalts (140). These calculations have considerable success in explaining boththe structure and stability of casein-coated emulsions, but they are lessadaptable to explain the behavior of more rigid protein surfactants.However, the same principles have been used to explain the apparentlyanomalous adsorption of phosvitin (61).

The difficulty of ascertaining the structure of the adsorbed protein isgreater for globular proteins. In these cases, the adsorbed layer is muchthinner than it is for the caseins, so that layers of b-lactoglobulin appearto be of the order of 1–2 nm thick instead of the value of about 10 nmmeasured for the caseins (83,104,137). Hydrodynamic and scattering experi-ments suggest that the thickness of the adsorbed layers is smaller than wouldbe expected from the protein in its natural conformation, so that thesesimple measurements of the size of the adsorbed protein already suggestthat adsorption causes a conformational change. This is confirmed byother techniques. For example, adsorbed b-lactoglobulin forms intermole-cular disulfide bonds (24), which does not occur when the molecules are intheir native conformations in solution (although the high concentration ofprotein in the adsorbed layer will certainly enhance any tendency that themolecules have to aggregate). In addition, detailed studies of the DSC of


emulsions containing b-lactoglobulin (33) have shown that (a) the proteinwhen adsorbed to the oil–water interface in emulsions loses its heat ofdenaturation (i.e., shows no intake of heat which can be associated withdenaturation, presumably because the protein is already surface denatured)and (b) if the protein is desorbed from the interface by treatment withdetergent (Tween-20), it can be seen to be denatured irreversibly (i.e., norecovery of the denaturation endotherm is seen). This may be contrastedwith the behavior of a-lactalbumin, which loses its heat of denaturationwhen adsorbed but recovers its original thermal behavior when the proteinis competitively desorbed by Tween (33,70). These studies confirm thatdifferent proteins show quite different degrees of denaturation when theyare adsorbed to oil–water interfaces.

Similarly, the use of infrared spectroscopy (FTIR) shows that signifi-cant changes in the conformation of b-lactoglobulin occur as a result ofadsorption. Detailed interpretation of the spectra is difficult, but changesin the contents of a-helix and b-sheet contents of the protein appear to occur(34). The protein a-lactalbumin is less affected by adsorption (71). However,nearly all of the evidence [chemical, physical, spectroscopic, and enzymolo-gical (32)] combines to show that adsorption-induced conformationalchanges occur. Moreover, these changes appear to be nearly always irrever-sible; that is, even if a protein is desorbed from the interface, it cannotrecover its original conformational state.

When considering the structure of a protein-based interfacial layer,there are other factors to be considered, namely the possibility that multi-layers, rather than monolayers, are formed and the possibility that specificproteins may exhibit variable behavior depending on the conditions. Caseinsare capable of this; for example, it is possible to prepare stable emulsionscontaining 20% (w/w) of soya oil, with as little as 0.3% casein, and in these,the surface coverage has been measured to be slightly less than 1mg/m2. Thehydrodynamic thickness of the adsorbed monolayer in these emulsions isabout 5 nm. In emulsions prepared with larger amounts of casein (1–2%),the surface coverage increases to 2–3mg/m2 and the thickness of theadsorbed monolayer is about 10 nm (i.e., about twice that of the layer atlower surface coverage) (94). It has been suggested that this is a result of theadsorbed casein molecules adopting two different conformations: one at lowcoverage, where the proteins have to cover a maximum area of surface(about 48 nm2 per molecule), and one at high coverage, where the moleculesare more closely packed (about 13 nm2 per molecule). The addition of moreprotein to the aqueous phase of emulsions made with low concentrations ofcaseinate results in the adsorption of some of the added protein and anincrease in the thickness of the adsorbed layers. This is true, even ifthe added protein is not casein, and illustrates that, for caseins at least,


the proteins on the surface possess sufficient mobility to be moved as otherproteins adsorb.

So far, we have considered monolayers. However, caseins and otherproteins can form multilayers; this has been demonstrated for adsorption toplanar interfaces, where large surface excesses are easily generated andmultilayers are formed (80). There is less evidence for this in emulsions,although some high surface coverages (up to about 10mg/m2) have beenmeasured (141), which can only arise from the presence of more than asingle layer on the oil–water interface because it would be impossible topack this amount of protein into a monomolecular layer. It is not clearwhy monolayers in some cases and multilayers in other cases are formed,although it is likely that the physical conditions of homogenization may beimportant, as well as there being a need for high concentrations of protein.Also, differences in the methods of preparing the caseins may be relevant; atneutral pH values, highly purified caseins have not generally been associatedwith multilayer formation, which seems generally to be associated with theuse of commercial sodium caseinates.

Multiple layers seem to be more easily formed in emulsions containingwhey proteins than with caseins. Because the whey proteins project less intosolution than do the caseins, they may be less effective than caseins atsterically preventing the approach of additional molecules that go to formthe multilayers. Finally, because they change conformation when theyadsorb, they may offer new possibilities for interaction with incomingwhey proteins from solution. There is evidence for multiple layers of wheyproteins from both planar interfaces and emulsion droplets (11,142).However, although these multiple layers exist, there is no definite evidencethat links them to changes in the functional properties of the emulsions. Noris it well determined how stable the multiple layers are, compared with amonolayer. It is very difficult, or perhaps impossible, to simply washadsorbed proteins from adsorbed monolayers formed on the interfaces ofoil droplets (143). However, the outer portion of multilayers may be morereadily displaced because it is held in place by protein–protein interactionsonly, which may be weaker than the forces which lead to adsorption.Generally, however, the properties of the outer parts of multilayers havebeen little studied.

C. Emulsions Stabilized by Particles

As a final degree of complexity, food emulsions may be stabilized by par-ticles. Perhaps the most common are the protein ‘‘granules’’ from egg yolk,which play a role in the stabilization of mayonnaise (89), and casein micellesin products such as homogenized milk and in ice creams. Both of these


emulsifiers are known to be adsorbed to the oil–water interface as complexparticles, which do not dissociate completely to their individual proteinseither during or after adsorption (144,145). During the homogenization ofmilk, casein micelles are partially disrupted at the oil–water interface so thatthey adsorb either whole or in fragments. Indeed, once a micelle hasadsorbed, it appears to be able to spread over an area of the interface(105,146,147). Thus, the fat droplets in homogenized milk are surroundedby a membrane that must contain some of the original fat globule mem-brane [phospholipid and protein (148)] but is primarily constituted of semi-intact casein micelles. If the milk has been heated before or after homoge-nization, whey proteins also form part of the layer surrounding the fatglobules (39,149). Likewise, the oil–water interface in mayonnaise is partlycoated by the granular particles formed from the phosphoprotein andlipoprotein constituents of egg yolk (150). In this high-lipid product, thegranules also may act to keep the oil droplets well separated and preventcoalescence.

Natural fat globule membrane was mentioned earlier as a possibleemulsifying agent (151). In its most native state (i.e., prepared in the labora-tory from unheated milk), this membrane is an effective emulsifier (152),although little is known of its structure as it resurrounds oil droplets.However, heating at temperatures greater than about 65�C is sufficient todenature the proteins of the membrane. This appears to greatly diminish theefficacy of the membrane material as an emulsifying agent (153).

Emulsion formation by means of these aggregates of protein is gen-erally less efficient than by the proteins when they are present in the mole-cular state, simply because the efficient formation of the emulsion dependson rapid coverage of the newly formed oil surface in the homogenizer. Aparticle containing many molecules of protein will encounter a fat surfaceless frequently than an equivalent amount of molecular protein. Thus,although it is possible to prepare homogenized milk with the proportionsof casein and fat which occur naturally in milk (a ratio of about 1:1.5 w/w),it is not possible to use 1% w/w of micellar casein to stabilize an emulsioncontaining 20% oil (154). On the other hand, 1% casein in a molecular form(sodium caseinate) is quite sufficient to form a finely dispersed, stable emul-sion with 20% oil. Therefore, unless the micelles are expected to confer somespecific advantage on the functional properties of the emulsion or unlessthere are specific legislative reasons, it is generally more effective to usecaseinate than casein micelles to stabilize an emulsion. With egg granules,the situation is different, inasmuch as the individual proteins from the egggranules are not readily available in a purified form analogous to caseinate.In this case, the choice between particulate and molecular forms of theprotein does not arise.


VII. MODIFICATION OF THE INTERFACIAL LAYER AFTEREMULSION FORMATION

In many food emulsions, more than one surfactant is present, so thatmixtures of proteins, small-molecule surfactants (oil soluble and watersoluble), and lecithins may be on the interface or in the continuous phase.Under these circumstances, the interfacial layer will contain more than onetype of molecule. The properties of the emulsion (the sizes of the droplets,the functionality, and the stability) will, in turn, depend on which types ofmolecule in the formulation are actually on the interface and what effectexternal factors have on the conformations of the adsorbed materials(Fig. 3).

A. Interfaces Containing Mixtures of Proteins

It has been shown that in mixtures of proteins in emulsions formed atneutral pH and moderate temperatures, there is generally no competitionfor the interface. For example, there is no preferential adsorption betweenthe proteins when a mixture of a-lactalbumin and b-lactoglobulin is homo-genized with oil; the amounts of protein adsorbed are in proportion to theirconcentrations (155). The same is true even when a mixture of sodium case-inate and whey protein is used as the surfactant in an emulsion (11). Theonly case where competitive adsorption has been truly observed is when

Figure 3 Examples of the potential modification of the adsorbed layer of protein

during storage or processing of an emulsion stabilized by a layer of sodium caseinate.

Various treatments (see text for details) can give rise to a number of modifications of

the surface, which, in turn, affect the stability properties of the emulsion. The list of

modifications is not exhaustive.


b-casein is used to displace adsorbed as1-casein, and vice versa, so that thereis a possibility that these two proteins adsorb according to thermodynamicequilibrium (12). Even this observation is complicated, however, because itappears to apply only to mixtures of highly purified caseins; the displace-ment reactions with commercial sodium caseinate (where a similar resultwould a priori be expected), give much less clear results (16,156).

Rather than forming an emulsion in the presence of mixed surfactants,it is possible to form an emulsion using one protein and then to attempt todisplace that protein from the interface with another. It is a general finding,however, that usually the protein which is first on the interface resists dis-placement (157). Because of its high surface activity and flexibility, b-caseinappears to be the best displacing agent, but it is not always capableof displacing an already adsorbed protein. It can displace as1-casein ora-lactalbumin from an interface (12,158), but the process is more complexwith adsorbed b-lactoglobulin (159), especially if the emulsion containingthe b-lactoglobulin has been allowed to age before the b-casein is added.

The difficulty of replacing one protein with another is perhaps notsurprising, because proteins are adsorbed to the interface by many indepen-dent points of contact (because they have several hydrophobic regions capa-ble of binding to an interface). For all of these contacts to become desorbedat once is extremely unlikely, and so the spontaneous desorption of a proteinmolecule is very rare; this is why it is difficult to wash proteins from an oil–water interface (143). Replacement of an adsorbed protein molecule by onefrom solution must presumably require a concerted movement of the twomolecules; as parts of one are displaced, they are replaced by parts of theother, until, finally, one of the two proteins is liberated into the bulk solu-tion. Even this process, although more likely than spontaneous desorption,is not certain to succeed, especially if the adsorbed protein has been on theinterface for some time and has been able to make bonds with neighboringmolecules. Moreover, given the very high concentration of protein in theadsorbed layer (see earlier text), it may even be difficult for a second type ofprotein to penetrate the adsorbed layer to initiate the displacement process.Therefore, although thermodynamic considerations may favor one proteinover another, kinetic factors militate against rapid exchange. Nonetheless,there do seem to be factors that influence the competition between proteins.As has been suggested earlier hydrophobicity and flexibility may be impor-tant criteria (because b-casein can be an effective displacing agent).Especially, whatever increases flexibility may lead to increasing competitive-ness. The most obvious example of such a change is a-lactalbumin; in itsnative state, this protein has a globular structure partly maintained by thepresence of one bound calcium ion (160). Removal of this Ca2þ leads to theprotein adopting a ‘‘molten globule’’ state, whose tertiary structure is


altered (161), and this leads to increased flexibility and competitiveness atthe interface (162). The removal of the Ca2þ can be achieved by chelation orby reducing the pH, and under these conditions, a-lactalbumin out-com-petes b-lactoglobulin for adsorption to the interface (102,104,162). To someextent, the competition can be reversed by reneutralizing in the presence ofCa2þ, so that in this case, there is dynamic competition between the proteinsfor the interface, not simply preferential adsorption during emulsion forma-tion (104).

It should be noted that competition between proteins need not occur.For example, lysozyme (being positively charged at neutral pH) formscomplexes with other egg-white proteins (negatively charged) on air–waterinterfaces. The complex may adsorb to the interface or the lysozyme maybind to already adsorbed proteins (163). This coadsorption is, however, notgeneral, because it requires two proteins to be of opposite charge, which isrelatively rare.

B. Addition of Small-Molecule Emulsifiers to a Protein-StabilizedEmulsion

Competition between adsorbed and free proteins can be considerablyenhanced by the presence of small surfactant molecules (164). In suchcases, there is competition between the proteins and small molecules aswell as between the proteins themselves. However, in such a case, insteadof the desorption of a protein requiring the inefficient process of simulta-neous detachment at all points, or the slow creeping displacement of oneprotein molecule by another, it is possible for a number of small molecules todisplace a protein by separately replacing the individual points of attach-ment. It is known that small-molecule surfactants are capable of efficientlydisplacing adsorbed proteins, although the details of the reactions depend onthe type of surfactant and whether it is oil or water soluble (165–168). Water-soluble surfactants are capable of removing all of the adsorbed protein fromthe oil–water interface, although they may require a molecular ratio of about30:1 surfactant:protein (164). At lower ratios, some, but not all, of the pro-tein is displaced (Fig. 4). This displacement occurs either when the small-molecule emulsifier is present at the moment of homogenization or it is addedlater. Oil-soluble surfactants (low HLB numbers) are less effective at com-pletely displacing protein or of preventing protein adsorption (16,168,169).For solubility reasons, these surfactants cannot be added to the emulsionafter it has been formed, but must be incorporated during the homogeniza-tion step. In addition to competing with proteins for adsorption to the oil–water interface, both during formation of the emulsion and its subsequentstorage, some small-molecule surfactants also facilitate the exchange


reactions of the proteins themselves. For example, although a-lactalbuminand b-lactoglobulin do not compete well with each other under normal cir-cumstances at neutral pH (155), the presence of Tween causes the adsorptionof a-lactalbumin to be favored over b-lactoglobulin (164). Presumably, thepresence of surfactant enables a more thermodynamic equilibrium to beestablished, rather than the extremely slow kinetically determined exchangewhich normally occurs (if it occurs at all) between the two proteins.Alternatively, if the surfactant actually binds to the protein, its conformationmay change so that it becomes more surface active, as was shown for the‘‘molten globule’’ conformation of a-lactalbumin. Similarly, the presence ofsurfactants can alter the exchange between b-casein and b-lactoglobulin onan oil–water interface. Of the two proteins, b-casein is displaced first by bothoil-soluble and water-soluble surfactants (18,170).

The displacement of proteins by small-molecule surfactants has beenstudied in some detail, using the technique of atomic force microscopy(AFM) which allows the detailed visualization of adsorbed films on a

Figure 4 The displacement of caseins from an emulsion by the incorporation of the

water-soluble surfactant Tween-20 (lower line) or the oil-soluble surfactant Span-20

(upper line). The water-soluble material can be added at any time before or after the

emulsion is formed and can give almost complete displacement of the proteins. The

oil-soluble surfactant must be present dispersed in the oil phase before emulsion

formation takes place.


planar interface. A series of reactions has been shown to occur whenproteins are displaced by Tween. Rather than the protein molecules beingdisplaced one by one, it is found that the surfactant pushes the protein asidefrom parts of the interface (171); in effect, a two-dimensional phaseseparation occurs. As this occurs, protein does not leave the surface, butforms multilayers over restricted areas of the interface (18). This ‘‘orogenic’’displacement reaction explains some anomalous research results on theapparent thickness of adsorbed layers (17) and also offers the interestingpossibility that the inhomogeneous structure of the interface may lead toareas of different reactivity. The results using AFM have been confirmed bystudies using neutron reflectivity (172), which showed that the removal ofthe protein (b-lactoglobulin) by the surfactant is not a single continuousprocess. Modeling of the displacement reaction has been able to reproducesome aspects of the exchange behavior, using a fairly simple model(173,174). What must be critical to the behavior is the respective influenceof interprotein and protein–solvent interactions. Because in practice theprotein is displaced from the interface in an aggregated form, the changesin the protein structures and properties arising from adsorption anddesorption can be qualitatively understood.

These observations lead to the question of whether the same type ofbehavior occurs when two proteins are present on an interface together.Here, the situation is not clear. Studies have been made on mixed proteinsat air–water interfaces, and whereas in one case a slow phase-separationbehavior has been seen over the course of some days (175,176), in anotherstudy by a different group, no such separation could be seen (18). It will benecessary to await further developments in this field before coming to adecision. Neither of the studies used emulsion droplets, which are not sus-ceptible to the techniques, but it is intriguing to contemplate the possibilityof producing controlled multifunctional surfaces by mixing appropriateproteins.

Lecithins represent a different type of small-molecule emulsifier.Although these molecules possess surfactant properties, they do notbehave like other small-molecule emulsifiers. For example, they do notappear to displace proteins efficiently from the interface, even though thelecithins may themselves become adsorbed (8). They certainly have the cap-ability to alter the conformation of adsorbed layers of caseins, although theway in which they do this is not fully clear; it is possibly because they can‘‘fill in’’ gaps between adsorbed protein molecules (177). In actual foodemulsions, the lecithins in many cases contain impurities, and the role ofthese (which may also be surfactants) may confuse the way that lecithin acts(178). It is possible also for the phospholipids to interact with the proteinpresent to form vesicles composed of protein and lecithin, independently of


the oil droplets in the emulsion. The existence of such vesicles has beendemonstrated (179), but their functional properties await elucidation.

C. Chemical Modification of the Interfacial Layer

Once a protein-stabilized emulsion has been formed, it is possible to modifythe interfacial layer by chemical reactions. In fact, spontaneous reactionsmay occur during processing and storage of an emulsion, which change thestructure and the properties of the interfacial layer. Both b-casein andb-lactoglobulin can suffer change in stored emulsions with soy oil (4,180).This may be attributed to the oxidation of the unsaturated fatty acid chainsof the oil giving rise to enal molecules, which subsequently react with lysineor arginine residues of the proteins.

Changes in the conformations of adsorbed proteins can be induced bychanges in the properties of the aqueous phase, especially for the caseins.Increases in ionic strength, or the presence of Ca2þ, cause the thickness ofthe adsorbed layers to decrease, because of smaller repulsive interactionsbetween the charges of the protein molecules (43,181). Similarly, the addi-tion of a poor solvent, such as ethanol, causes the adsorbed layer to collapseand lose its steric stabilizing properties (182). It is this effect that makescream liqueurs so susceptible to the presence of Ca2þ (183).

The description in the previous section of the only moderate exchangebetween adsorbed and free proteins refers to results at room temperature,which is where nearly all of the detailed studies have been made. However, itappears that the exchange is temperature dependent. It has been demon-strated recently that whey proteins (especially b-lactoglobulin) can displaceas1- and b-caseins from an oil–water interface during heating (this does notoccur at room temperature) (129). If whey protein isolate is added to anemulsion prepared using oil and sodium caseinate and the mixture is heatedto a temperature in excess of about 50�C, the whey proteins rapidly becomeadsorbed (Fig. 5), and as they do so, the caseins are desorbed, so the surfacecoverage by protein remains approximately constant (184). Only the majorcaseins are desorbed, however, with the as2- and k-caseins remaining on theemulsion droplets. This behavior has been insufficiently studied to be fullyunderstood. Interestingly, it is the whey protein which displaces b-casein;this is similar to the effect of surfactants described in the previous section. Itis not known why the minor caseins resist displacement; the obvious possi-bility, that disulfide bonds are formed between these caseins and the wheyproteins, does not seem to occur (185). It seems evident, however, that theexchange of proteins may occur more readily than was previously thought,especially in food preparations where an emulsion is added to a solutioncontaining other proteins and the mixture undergoes a heat treatment.


Mention has already been made of the disulfide-mediated polymeriza-tion of whey proteins (24). Evidently, the conformational changes producedas a result of adsorption allow the single sulfhydryl group of the protein tobecome accessible for reaction; normally, it is buried in the structure of theprotein. The details of the reaction are not elucidated; evidently, the ‘‘acti-vated’’ lactoglobulin molecules can react not only with one another but witha-lactalbumin as well, if it is present (74). It is even possible that the reactiveadsorbed molecules can pull in molecules from the aqueous phase and thatdisulfide bonds can be formed between proteins on different particles.However, the free-sulfhydryl groups are not universally accessible, becauseb-lactoglobulin bound to an emulsion was not capable of reacting withk-casein from caseinates added after the emulsion was formed, or frompure k-casein added subsequently (184,185).

Figure 5 Kinetics of the exchange of whey protein with caseins at an emulsion

interface. The figure shows the time-dependent adsorption of b-lactoglobulin from a

whey protein isolate solution added to an emulsion originally made with sodium

caseinate. Curves for the removal of the caseins from the surface are analogous. The

results shown are for 45�C (g), 50�C (œ), and 80�C (^). The results show that not

only is the rate of exchange temperature dependent, but so is the extent of

replacement of one protein by another. The transfer at 80�C is too rapid to measure

by the techniques employed in the research.


The sulfhydryl–disulfide exchange is a spontaneous phenomenon. It isalso possible, however, to use cross-linking enzymes, such as transgluta-minase, to polymerize the interfacial proteins (26). This polymerization, asmay be expected, produces strong, rigid, interfacial layers (186), as long asthe extent of cross-linking is not too large; these layers protect againstcoalescence and promote ethanol stability (187). Similar to the disulfide-induced polymerization, the cross-linking can be carried to an extent thatcauses the emulsions to destabilize (aggregate), presumably becausecross-links begin to form between protein molecules adsorbed to differentparticles.

Finally, in this section, we mention the effects of ultra high pressure onsystems containing emulsions. It seems probable that high pressure has aneffect on the behavior of the proteins in emulsions and, consequently,the exchange between them. Thus, unlike the effect of heat, where b-lacto-globulin was shown to displace the major caseins, the application of pres-sure in such mixed-protein emulsion systems favors the adsorption ofcaseins (188), possibly because the high pressure denatures the b-lactoglo-bulin which is in solution, causes it to aggregate, and renders it incapable ofadsorbing efficiently. Similar effects are found if a small-molecule emulsifieris mixed with the b-lactoglobulin-stabilized emulsion before pressure isapplied. Further evidence of the effect of pressure-induced changes in b-lactoglobulin emulsions comes from the increased interaction of the proteinwith pectin as a result of the high-pressure treatment. Even at neutral pH, aninteraction is observed (189). Clearly, the effect of pressure as a denaturingagent may influence the behavior even of adsorbed species.

It is worth noting that the structures of denatured proteins are differ-ent, depending on how they are denatured—whether by heat, by pressure, orby adsorption. The effects of combinations of these factors on the structuresand reactivities of proteins are largely unknown; in fact, the effects ofprocessing on emulsions are only now beginning to be studied in realisticsystems, a trend which is much to be encouraged.

VIII. CONCLUDING REMARKS

The object of making food emulsions is to provide a stable and controllablesource of food, whose texture, taste, and nutritional and storage propertiesare acceptable to the consumer. Although the number of possible ingredi-ents is limited by the constraints of healthy nutrition, it is neverthelessevident that within the available range, there is a great deal of opportunityfor variation in the properties of the emulsions—for instance, the particlesize and the composition of the stabilizing layer of the interface, which, in


turn, influence the stability and functional behavior of the emulsion. On theother hand, many emulsions used in foods have their roots in establishedformulations, and an understanding of why certain emulsions behave asthey do is still not established in a number of cases.

In the manufacture of many real food emulsions, the path followed iscritical, emphasizing once again that emulsions, like many other food sys-tems, are not in equilibrium states and that two products may be verydifferent although they have the same overall composition. On this point,our knowledge is insufficient and needs to be extended. For example, theheat or high-pressure treatment of ingredient proteins either before or afterthe formation of an emulsion may critically affect the behavior of the emul-sion. As foods containing emulsified material become more complex orsophisticated in their ingredients, the level of understanding required tocontrol their formation and properties is increased. The challenge for thefuture is to be able to describe and control some of the most complexemulsions so as to enable greater functional stability for these food systems.

A further aspect, which is becoming of ever-increasing importance inthe public mind, is that of the nutritional function of food emulsions.Reduced-fat formulations of traditional products are demanded, which,nevertheless, are required to possess textural and organoleptic propertiesas close as possible to those of the traditional material. This in itself providesa challenge to the emulsion technologist—to reproduce the properties whilereducing the amount of ‘‘active’’ constituent. The demand for the incorpora-tion of nutritionally beneficial lipid materials (phospholipids, specific fattyacids) also produces a challenge; the incorporation of these materials intofoods, complete with antioxidants and other necessary ingredients, willrequire increased ingenuity on the part of the emulsion technologist. Inaddition, there is discussion of targeting the materials contained in afood. No longer is it sufficient simply to provide nutrition; ideally, it isnecessary to define in which portion of the digestive tract the componentsof the food are to be liberated. Already there are encapsulated materialsavailable whose coatings are designed for this purpose. With emulsions ofspecific oils being part of the ‘‘functional food’’ system, we may expect to seeincreased demand for emulsions which are controlled not only during man-ufacture but during consumption as well. This represents a real challenge forthe emulsion technologist of the future.

REFERENCES

1. H. D. Goff, J. Dairy Sci. 80, 2620–2630 (1997).

2. H. D. Goff, Int. Dairy J. 7, 363–373 (1997).


3. J. M. Aguilera and H. G. Kessler, J. Food Sci. 54, 1213–1221 (1989).

4. E. M. Stevenson, D. S. Horne and J. Leaver, Food Hydrocolloids 11, 3–6 (1997).

5. H. D. Goff, M. Liboff, W. K. Jordan, and J. Kinsella, Food Microstruct. 6,

193–198, (1987).

6. D. F. Darling, J. Dairy Res. 49, 695–712 (1982).

7. B. E. Brooker, Food Strut. 9, 223–230 (1990).

8. J.-L. Courthaudon, E. Dickinson, and W. W. Christie, J. Agric. Food Chem. 3,

1365–1368 (1991).

9. S. O. Agboola, H. Singh, P. A. Munro, D. G. Dalgleish, and A. M. Singh,

J. Agric. Food Chem. 46, 1814–1819 (1998).

10. D. G. Dalgleish, Trends Food Sci. Technol. 8, 1–6 (1997).

11. J. A. Hunt and D. G. Dalgleish, Food Hydrocolloids 8, 175–187 (1994).

12. E. Dickinson, S. E. Rolfe, and D. G. Dalgleish, Food Hydrocolloids 2, 397–405

(1988).

13. H. S. Rollema, in Advanced Dairy Chemistry, 1. Proteins (P. F. Fox, ed.,)

Elsevier Science, Barking, UK, 1992, p. 111–140.

14. M. Srinivasan, H. Singh, and P. A. Munro, Int. Dairy J. 9, 337–341, 1999.

15. J. Chen and E. Dickinson, J. Sci. Food Agric. 62, 283–289 (1993).

16. S. E. Euston, H. Singh, D. G. Dalgleish, and P. A. Munro, J. Food Sci., 60,

1124–1131 (1995).

17. D. G. Dalgleish, M. Srinivasan, and H. Singh, J. Agric. Food Chem., 43,

2351–2355 (1995).

18. A. R. Mackie, A. P. Gunning, M. Ridout, P. Wilde, and V. J. Morris, Langmuir

17, 6593–6598 (2001).

19. P. W. J. R. Caessens, S. Visser, H. Gruppen, G. A. van Aken, and A. G. J.

Voragen, Int. Dairy J. 9, 347–351 (1999).

20. P. W. J. R. Caessens, H. Gruppen, C. J. Slangen, S. Visser, and A. G. J.

Voragen, J. Agric. Food Chem. 47, 1856–1862 (1999).

21. X. L. Huang, G. L. Catignani, and H. E. Swaisgood, J. Agric. Food Chem. 44,

3437–3443 (1996).

22. W. U. Wu, N. S. Hettiarachchy, and M. Qi, J. Am. Oil Chem. Soc. 75, 845–850

(1998).

23. M. Qi, N. S. Hettiarachchy, and U. Kalapathy, J. Food Sci. 62, 1110–1115

(1997).

24. E. Dickinson and Y. Matsumura, Int. J. Biol. Macromol. 13, 26–30 (1991).

25. S. Roth, B. S. Murray, and E. Dickinson, J. Agric. Food Chem. 48, 1491–1497

(2000).

26. M. Færgemand and B. S. Murray, J. Agric. Food Chem. 46, 884–890 (1998).

27. D. G. Dalgleish, Colloids Surfaces B 1, 1–8, (1993).

28. T. Nylander and M. Wahlgren, J. Colloid Interf. Sci. 162, 151–162 (1994).

29. T. Nylander, F. Tiberg, and M. Wahlgren, Int. Dairy J. 9, 313–317 (1999).

30. E. Dickinson, D. S. Horne, J. Phipps, and R. Richardson, Langmuir 9, 242–248

(1993).

31. F. A. Husband, M. J. Garrood, A. R. Mackie, G. R. Burnett, and P. J. Wilde,

J. Agric. Food Chem. 49, 859–866 (2001).


32. A. L. de Roos and P. Walstra, Colloids Surfaces B 6, 201–208 (1996).

33. M. Corredig and D. G. Dalgleish, Colloids Surfaces B 4, 411–422 (1995).

34. Y. Fang and D. G. Dalgleish, J. Colloid Interf. Sci. 196, 292–298 (1998).

35. S. O. Agboola and D. G. Dalgleish, J. Agric. Food Chem. 44, 3637–3642

(1996).

36. T. van Vliet, Colloid Polym. Sci. 266, 518–524 (1988).

37. C. K. Reiffers-Magnani, J.-L. Cuq, and H. J. Watzke, Food Hydrocolloids 14,

521–530 (2000).

38. R. Jost, F. Dannenberg, and J. Rosset, Food Microstruct. 8, 23–28 (1989).

39. S. K. Sharma and D. G. Dalgleish, J. Dairy Res. 61, 375–384 (1994).

40. H. Casanova and E. Dickinson, J. Agric. Food Chem. 46, 72–76 (1998).

41. E. Dickenson and E. Davies, Colloids Surfaces B 12, 203–212 (1999).

42. E. P. Schokker and D. G. Dalgleish, J. Agric. Food Chem. 48, 198–203 (2000).

43. D. V. Brooksbank, C. M. Davidson, D. S. Horne, and J. Leaver, J. Chem. Soc.

Faraday Trans. 89, 3419–3425 (1993).

44. E. Dickinson, M. I. Goller, and D. J. Wedlock, J. Colloid Interf. Sci. 172,

192–202 (1995).

45. E. Dickinson and M. Golding, Food Hydrocolloids 11, 13–18 (1997).

46. D. Reichman and N. Garti in Food Polymers, Gels and Colloids (E. Dickinson,

ed.), Royal Society of Chemistry, Cambridge, 1991, pp. 549–556.

47. D. M. W. Anderson, in Gums and Stabilisers for the Food Industry 3 (G. O.

Phillips, D. J. Wedlock, and P. A. Williams eds.), Elsevier Applied Sciences,

Barking, UK, 1986, 79–86.

48. B. F. M. Namee, E. D. O’Riordan, and M. O’Sullivan, J. Agric. Food Chem. 46,

4551–4555 (1998).

49. Y. Matsumura, C. Satake, M. Egami, and T. Mori, Biosci. Biotechnol. Biochem.

64, 1827–1835 (2000).

50. E. Dickinson and V. B. Galazka, Food Hydrocolloids 5, 281–296 (1991).

51. M. Darewicz and J. Dziuba, Nahrung 45, 15–20 (2001).

52. D. G. Dalgleish and A.-L. Hollocou, in Food Colloids—Proteins, Lipids and

Polysaccharides (E. Dickenson and B. Bergenstahl, eds.), Royal Society of

Chemistry, Cambridge, 1997, pp. 236–244.

53. E. Dickinson, M. G. Semenova, A. S. Antipova, and E. G. Pelan, Food

Hydrocolloids, 12, 425–432 (1998).

54. D. F. Darling and R. J. Birkett, in Food Emulsions and Foams (E. Dickinson,


55. N. Krog, in Microemulsions and Emulsions in Foods (M. El-Nokaly

and D. Cornell eds.), American Chemical Society, Washington, DC, 1991,

pp. 139–145.

56. E. Li-Chan and S. Nakai, in Food Proteins (J. E. Kinsella and W. G. Soucie,

eds.), American Oil Chemists’ Society, Champaign, IL, 1989, pp. 232–251.

57. E. Dickinson, D. J. Pogson, E. W. Robson, and G. Stainsby, Colloids Surfaces

14, 135–141 (1985).

58. A. Kato, N. Tsutsui, N. Matsudomi, K. Kobayashi, and S. Nakai, Agric. Biol.

Chem. 45, 2755–2760 (1981).


59. E. Dickinson, J. A. Hunt, and D. G. Dalgleish, Food Hydrocolloids 4, 403–414

(1991).

60. A. Kato, S. Miyazaki, A. Kawamoto, and K. Kobayashi, Agric. Biol. Chem. 51,

2989–2994 (1987).

61. E. Dickinson, V. J. Pinfield, and D. S. Horne, J. Colloid Interf. Sci. 187,

539–541 (1997).

62. S. Damodaran and S. Xu, J. Colloid Interf. Sci. 178, 426–435 (1996).

63. B. M. Byrne, A. D. van het Schip, J. A. M. van de Klundert, A. C. Arnberg,

M. Gruber, and G. Ab, Biochemistry 23, 4275–4279 (1984).

64. E. Keshavarz and S. Nakai, Biochim. Biophys. Acta 576, 269–279 (1979).

65. S. Nakai, J. Agric. Food Chem. 31, 676–683 (1983).

66. A. Kato and S. Nakai, Biochim. Biophys. Acta 624, 13–20 (1980).

67. F. MacRitchie and N. F. Owens, J. Colloid Interf. Sci. 29, 66–71 (1969).

68. C. A. Haynes and W. Norde, Colloids Surfaces B 2, 517–566 (1994).

69. W. Norde, F. MacRitchie, G. Nowicka, and J. Lyklema, J. Colloid Interf. Sci.

112, 447–456 (1986).

70. R. J. Green, I. Hopkinson, and R. A. L. Jones in Food Emulsions and Foams

(E. Dickinson and J. M. Rodriguez Patino, eds.), Royal Society of Chemistry,

Cambridge, 1999, pp. 285–295.

71. Y. Fang and D. G. Dalgleish, Food Hydrocolloids 12, 121–126 (1998).

72. W. Norde and J. P. Favier, Colloids Surfaces 64, 87–93 (1992).

73. M. C. L. Maste, E. H. W. Pape, A. van Hoek, W. Norde, and A. J. W. G.

Visser, J. Colloid Interf. Sci., 180, 632–633 (1996).

74. S. Damodaran and K. Anand, J. Agric. Food Chem. 45, 3813–3820 (1997).

75. E. A. Foegeding, in Food Proteins (J. E. Kinsella and W. G. Soucie, eds.),

American Oil Chemists’ Society, Champaign, IL, 1989, pp. 185–194.

76. K. D. Caldwell, J. Li, J.-T. Li, and D. G. Dalgleish, J. Chromatogr. 604, 63–71

(1992).

77. C. A. Haynes, E. Sliwinsky, and W. Norde, J. Colloid Interf. Sci. 164, 394–409

(1994).

78. C. Holt and L. Sawyer, Protein Eng., 2, 251–259 (1988).

79. H. E. Swaisgood, in Advanced Dairy Chemistry, 1. Proteins (P. F. Fox, ed.),

Elsevier Science, Barking, UK, 1992, pp. 63–110.

80. D. E. Graham and M. C. Phillips, J. Colloid Interf. Sci. 70, 415–426 (1979).

81. M. C. Phillips, Chem. Ind. 170–176 (1977).

82. D. E. Graham and M. C. Phillips, J. Colloid Interf. Sci. 70, 427–439 (1979).

83. D. G. Dalgleish and J. Leaver, in Food Polymers. Gels and Colloids

(E. Dickinson, ed.), Royal Society of Chemistry, Cambridge, 1991, pp. 113–122.

84. S. G. Hambling, A. S. M. Alpine, and L. Sawyer, in Advanced Dairy Chemistry,

1. Proteins (P. F. Fox, ed.), Elsevier Science, Barking, UK, 1992, pp. 141–190.

85. R. D. Bee, J. Hoogland, and R. H. Ottewill, in Food Colloids and Polymers,

Stability and Mechanical Properties (E. Dickinson and P. Walstra, eds.), Royal

Society of Chemistry, Cambridge, 1993, pp. 341–353.

86. Y. Kamata, K. Ochiai, and F. Yamauchi, Agric. Biol. Chem. 48, 1147–1152

(1984).


87. P. Plietz, G. Damaschun, J. J. Muller, and K.-D. Schwenke. Eur. J. Biochem.

130, 315–320 (1983).

88. D. Causeret, E. Matringe, and D. Lorient, J. Food Sci. 56, 1532–1536 (1991).

89. L. D. Ford, R. Borwankar, R. W. Martin, Jr., and N. Holcomb, in Food

Emulsions 3rd ed. (K. Larson and S. E. Friberg, eds.), Marcel Dekker,

New York, 1997, pp. 361–412.

90. P. Walstra, in Food Structure and Behavior (J. M. V. Blanshard and P. Lillford,

eds.), Academic Press, London, 1987, pp. 87–106.

91. P. Walstra, in Encyclopedia of Emulsion Technology (P. Becher, ed.), Marcel

Dekker, New York, 1983, pp. 57–128.

92. L. Taisne, P. Walstra, and B. Cabane, J. Colloid Interf. Sci. 184, 378–390

(1996).

93. O. Robin, N. Remillard, and P. Paquin. Colloids Surfaces A, 80, 211–222

(1993).

94. Y. Fang and D. G. Dalgleish, J. Colloid Interf. Sci. 156, 329–334 (1993).

95. E. Dickinson, F. O. Flint, and J. A. Hunt, Food Hydrocolloids 3, 389–397

(1989).

96. L. V. Ogden, P. Walstra, and H. A. Morris, J. Dairy Sci. 59, 1727–1737 (1976).

97. C. Holt, Adv. Protein Chem. 43, 63–151 (1992).

98. H. Oortwijn, P. Walstra, and H. Mulder, Netherlands Milk Dairy J. 31,

134–147 (1977).

99. P. Walstra, Netherlands Milk Dairy J. 34, 181–190 (1980).

100. S. R. Euston and R. L. Hirst, J. Food Sci. 65, 934–940 (2000).

101. H. Pessen, J. M. Purcell, and H. M. Farell, Biochim. Biophys. Acta 828, 1–12

(1985).

102. J. A. Hunt and D. G. Dalgleish, J. Agric. Food Chem. 42, 2131–2135 (1994).

103. S. L. Turgeon, C. Sanchez, S. F. Gauthier, and P. Paquin, Int. Dairy J. 6,

645–608 (1996).

104. J. A. Hunt and D. G. Dalgleish, Food Hydrocolloids 10, 159–165 (1996).

105. W. Buchheim and P. Dejmek, in Food Emulsions, 3rd ed., (K. Larsson and

S. E. Friberg, eds.), Marcel Dekker, New York, 1997, pp. 235–278.

106. T. Gershank, S. Benzeno, and S. Benita, Pharmaceut. Res. 15, 863–868 (1998).

107. K. N. Pearce and J. E. Kinsella, J. Agric. Food Chem. 26, 716–723 (1978).

108. P. Walstra, J. Colloid Interf. Sci. 27, 493–500 (1968).

109. K. Strawbridge and F. R. Hallet, Macromolecules 27, 2283–2290 (1994).

110. K. Strawbridge and J. Watton, Can. J. Spectrosc. 36, 53–60 (1991).

111. K. Strawbridge and F. R. Hallet, Can. J. Phys. 70, 401–406 (1992).

112. J.-L. Gelin, L. Poyen, J.-L. Courthaudon, M. Le Meste, and D. Lorient, Food

Hydrocolloids 8, 299–308 (1991).

113. D. G. Dalgleish and F. R. Hallett, Food Res. Int. 28, 181–193 (1995).

114. F. R. Hallett, T. Craig, J. Marsh, and B. Nickel, Can. J. Spectrosc. 34, 63–70

(1989).

115. C. Urban and P. Schurtenberger, J. Colloid Interf. Sci. 207, 150–158 (1998).

116. C. Urban and P. Schurtenberger, Phys. Chem. Chem. Phys. 1, 3911–3915

(1999).


117. D. S. Horne and C. Davidson, Colloids Surfaces A 77, 1–8 (1993).

118. D. G. Dalgleish and D. S. Horne, Milchwissenschaft 46, 417–422 (1991).

119. A. Lips, T. Westbury, P. M. Hart, I. D. Evans, and I. J. Campbell, in Food

Colloids and Polymers, Stability and Mechanical Properties (E. Dickinson and

P. Walstra, eds.), Royal Society of Chemistry, Cambridge, 1993, pp. 31–44.

120. D. J. McClements, Langmuir 12, 3454–3461 (1996).

121. D. J. McClements and J. N. Coupland, Colloids Surfaces A 117, 161–170

(1996).

122. J. N. Coupland and D. J. McClements, J. Food Eng. 50, 117–120 (2001).

123. J. N. Coupland and D. J. McClements, J. Am. Oil Chem. Soc. 74, 1559–1564

(1997).

124. A. S. Dukhin and P. S. Goetz, Colloids Surfaces A 144, 49–58 (1998).

125. T. Wade and J. K. Beattie, Colloids Surfaces B, 10, 73–85 (1995).

126. L. Kong, J. K. Beattie, and R. J. Hunter, J. Colloid Interf. Sci. 238, 70–79

(2001).

127. J. M. M. Westerbeek and A. Prins, in Microemulsions and Emulsions in Foods

(M. El-Nokaly and D. Cornell, eds.), American Chemical Society,

Washington, DC, 1991, pp. 146–160.

128. A. Kabalnov, T. Tarara, R. Arlauskas, and J. Weers, J. Colloid Interf. Sci.

184, 227–235 (1996).

129. J. M. Brun and D. G. Dalgleish, Int. Dairy J. 9, 323–327 (1999).

130. F. MacRitchie, Colloids Surfaces A 76, 159–166 (1993).

131. J. Leaver and D. G. Dalgleish, Biochim. Biophys. Acta 1041, 217–222 (1990).

132. J. Leaver and D. G. Dalgleish, J. Colloid Interf. Sci. 149, 49–55 (1992).

133. D. G. Dalgleish, Colloids Surfaces 46, 141–155 (1990).

134. A. R. Mackie, J. Mingins, and A. N. North, J. Chem. Soc. Faraday Trans. 87,

3043–3049 (1991).

135. D. G. Dalgleish and J. Leaver, J. Colloid Interf. Sci. 141, 288–294 (1991).

136. M. Shimizu, A. Ametani, S. Kaminogawa, and K. Yamauchi, Biochim.

Biophys. Acta 869, 259–264 (1986).

137. A. R. Mackie, J. Mingins, and R. Dann, in Food Polymers, Gels and

Colloids (E. Dickinson, ed.), Royal Society of Chemistry, Cambridge, 1991,

pp. 96–111.

138. F. A. M. Leermakers, P. J. Atkinson, E. Dickinson, and D. S. Horne,

J. Colloid Interf. Sci. 178, 681–693 (1996).

139. E. Dickinson, D. S. Horne, V. J. Pinfield, and F. A. M. Leermakers, J. Chem.

Soc. Faraday Trans. 93, 425–432 (1997).

140. E. Dickinson, V. J. Pinfield, D. S. Horne, and F. A. M. Leermakers, J. Chem.

Soc. Faraday Trans. 93, 1785–1790 (1997).

141. M. Britten and H. J. Giroux, J. Agric. Food Chem. 41, 1187–1191 (1993).

142. M. Rosenberg and S. L. Lee, Food Struct. 12, 267–274 (1993).

143. F. MacRitchie, J. Colloid Interf. Sci. 105, 119–123 (1985).

144. D. G. Dalgleish and E. W. Robson, J. Dairy Res. 52, 539–546 (1985).

145. H. Oortwijn and P. Walstra, Netherlands Milk Dairy J. 33, 134–154 (1979).

146. P. Walstra and H. Oortwijn, Netherlands Milk Dairy J. 36, 103–113 (1982).


147. D. G. Dalgleish, S. Tosh, and S. West, Netherlands Milk Dairy J. 50, 135–148

(1996).

148. A. V. McPherson and B. J. Kitchen, J. Dairy Res. 50, 107–133 (1983).

149. M. E. Cano-Ruiz and R. L. Righter, J. Dairy Sci. 80, 2732–2739 (1997).

150. M. A. Tung and L. J. Jones, Scanning Electron Microsc. 3, 523–530 (1981).

151. C. Kanno, Y. Shimomura, and E. Takano, J. Food Sci. 56, 1219–1223 (1991).

152. M. Corredig and D. G. Dalgleish, J. Agric. Food Chem. 46, 2533–2540 (1998).

153. M. Corredig and D. G. Dalgleish, J. Dairy Res. 65, 465–477 (1988).

154. J. A. Hunt and D. G. Dalgleish, unpublished results.

155. E. Dickinson, S. E. Rolfe, and D. G. Dalgleish, Food Hydrocolloids 3, 193–203

(1989).

156. E. W. Robson and D. G. Dalgleish, J. Food Sci. 52, 1694–1698 (1987).

157. E. Dickinson, J. Food Eng. 22, 59–74 (1994).

158. E. Dickinson, S. E. Rolfe, and D. G. Dalgleish, Int. J. Biol. Macromol. 12,

189–194 (1990).

159. D. G. Dalgleish, S. E. Euston, J. A. Hunt, and E. Dickinson, in Food

Polymers, Gels and Colloids (E. Dickinson, ed.), Royal Society of Chemistry,

Cambridge, 1991, pp. 485–489.

160. M. Ikeguchi, K. Kuwajima, M. Mitani, and S. Sugai, Biochemistry 25,

6965–6972 (1986).

161. O. B. Ptitsyn, in Protein Folding (T. E. Creighton, eds.), W. H. Freeman,

New York, 1992, pp. 243.

162. E. Dickinson and Y. Matsumura, Colloids Surfaces B 3, 1–17 (1994).

163. S. Damodaran, K. Anand, and L. Razumovsky, J. Agric. Food Chem. 46,

872–876 (1998).

164. J.-L. Courthaudon, E. Dickinson, Y. Matsumura, and A. Williams, Food

Struct. 10, 109–115 (1991).

165. J.-L. Courthaudon, E. Dickinson, and D. G. Dalgleish, J. Colloid Interf. Sci.

145, 390–395 (1991).

166. J.-L. Courthaudon, E. Dickinson, Y. Matsumura, and D. C. Clark, Colloids

Surfaces 56, 293–300 (1991).

167. J. A. de Feijter, J. Benjamins, and M. Tamboer, Colloids Surfaces 27, 243–266

(1987).

168. E. Dickinson and S. Tanai, J. Agric. Food Chem. 40, 179–183 (1992).

169. E. Dickinson, R. K. Owusu, S. Tan, and A. Williams, J. Food Sci. 58, 295–298

(1993).

170. M. Cornec, P. J. Wilde, P. A. Gunning, A. R. Mackie, F. A. Husband, M. L.

Parker, and D. C. Clark, J. Food Sci. 63, 39–43 (1998).

171. A. P. Gunning, A. R. Mackie, A. R. Kirby, and V. J. Morris, Langmuir 17,

2013–2018 (2000).

172. D. S. Horne, P. J. Atkinson, E. Dickinson, V. J. Pinfield, and R. M.

Richardson, Int. Dairy J. 8, 73–77 (1998).

173. C. M. Wijmans and E. Dickinson, Phys. Chem. Chem. Phys. 1, 2141–2147

(1999).

174. C. M. Wijmans and E. Dickinson, Langmuir 15, 8344–8348 (1999).


175. L. Razumovsky and S. Damodaran, J. Agric. Food Chem. 49, 3080–3086

(2001).

176. T. Sengupta and S. Damodaran, J. Agric. Food Chem. 49, 3087–3091 (2001).

177. Y. Fang and D. G. Dalgleish, Colloids Surfaces B, 1, 357–364 (1993).

178. C. H. McCrae and D. D. Muir, J. Dairy Res. 59, 177–185 (1992).

179. Y. Fang and D. G. Dalgleish, Langmuir 11, 75–79 (1995).

180. J. Leaver, A. J. R. Law, E. Y. Brechany, and C. H. McCrae, Int. J. Food Sci.

Technol. 34, 503–508 (1999).

181. E. Dickinson, J. A. Hunt, and D. S. Horne, Food Hydrocolloids 6, 359–370

(1992).

182. D. S. Horne, Biopolymers 23, 989–993 (1984).

183. D. S. Horne and D. D. Muir, J. Dairy Sci. 73, 3613–3626 (1990).

184. D. G. Dalgleish, H. D. Goff, J. M. Brun, and B. Luan, Food Hydrocolloids 16,

303–311 (2002).

185. D. G. Dalgleish, H. D. Goff, and B. Luan, Food Hydrocolloids 16, 295–302

(2002).

186. M. Faergemand, B. S. Murray, and E. Dickinson, J. Agric. Food Chem. 45,

2514–2519 (1997).

187. M. Faergemand, J. Otte, and K. B. Qvist, Int. Dairy J. 8, 715–723 (1998).

188. E. Dickinson and J. D. James, J. Agric. Food Chem. 47, 25–30 (1999).

189. E. Dickinson and J. D. James, Food Hydrocolloids 14, 365–376 (2000).


2Food Emulsifiers: Their Chemicaland Physical Properties

Niels J. Krog and Flemming Vang SparsøDanisco, Brabrand, Denmark

I. INTRODUCTION

Foods are very complex colloidal systems and modern industrial pro-duction requires surface-active lipids such as emulsifiers as processingaids and to secure a uniform quality, improved texture, and longshelf life.

An emulsifier is defined as a substance that reduce surface tensionbetween oil–water or air–water, thus enhancing emulsification and increas-ing emulsion stability. Many native polar lipids and proteins comply withthis definition. Food emulsifiers, on the other hand, do not affect the emul-sification process significantly, but have other functions which are related tointerfacial properties and affect emulsion stability or destabilization ofwhippable emulsions. In addition, food emulsifiers have other functions infoods, such as the modification of fat crystallization, interactions with car-bohydrate components, or act as film forming substances, controllingoxygen or humidity transport, and such applications are not related to theclassical definition of an emulsifier.

The main goal of this chapter is to describe the relationship betweenthe chemical structure and composition of the emulsifiers and their func-tion in foods. The crystalline properties, interfacial adsorption, and inter-actions with food ingredients, such as fats, proteins, carbohydrates, andwater, are playing important roles in the function of emulsifiers in foodproducts.


II. CHEMICAL COMPOSITION AND PHYSICAL PROPERTIES

A. Monoglycerides

An overview of food emulsifiers and their identification numbers is pre-sented in Table 1. The use of mono-diglycerides dates back to the 1930s,when they were first used in margarine production.

Table 1 Food Emulsifiers and Their Legal Status

Chemical name Abbreviation ADI valuea EU No.

US FDA

21 CFR

Lecithin — Not limited E322 §184.1400b

Mono-diglycerides

(distilled monoglycerides) MAG Not limited E471 §184.1505b

Acetic acid esters of

monoglycerides ACETEM Not limited E472a §172.828

Lactic acid esters of

mono-diglycerides LACTEM Not limited E472b §172.852

Citric acid esters of

mono-diglyceries CITREM Not limited E472c §172.832

Diacetyl tartaric acid esters

of monoglycerides DATEM 0–50 E472e §184.1101b

Succinic acid esters of

monoglycerides SMG — — §172.830

Salts of fatty acids (Na, K, Ca) — Not limited E470a §172.863

Polyglycerol esters

of fatty acids PGE 0–25 E475 §172.854

Polyglycerol polyricinoleate PGPR 0–7.5 E476 —

Propylene gycol esters

of fatty acids PGMS 0–25c E477 §172.856

Sodium stearoyl-lactylate SSL 0–20 E482 §172.844

Calcium stearoyl-lactylate CSL 0–20 E481 §172.846

Sucrose esters of fatty acids — 0–10 E473 §172.859

Sorbitan monostearate SMS 0–25 E491 §172.842

Sorbitan tristearate STS 0–15 E492 —d

Polysorbate 60 PS 60 0–25 E435 §172.836

Polysorbate 65 PS 65 0–25 E436 §172.838

Polysorbate 80 PS 80 0–25 E433 §172.840

aAcceptable daily intake in mg/kg body weight per day.bGenerally recognized as safe (GRAS).cCalculated as propylene glycol.dPetition filed and accepted.


Monoglycerides are produced industrially by interesterification (glyc-erolosis) of edible fats or oils with glycerol, the reaction of the componentstakes place at high temperature (200–260�C) under alkaline catalysis. Thereaction scheme is shown in Fig. 1, and has been described in detail byFeuge and Bailey (1) and Sonntag (2). Table 2 shows the calculated com-position of the fat-phase mixture from the glycerolosis of fat with differentconcentrations of glycerol. The so-called monodiglycerides produced by

Table 2 Composition of Glycerolysis Equilibrium Mixtures

Amount of glycerol

added to triglyceride

(% w/w)

Equilibrium mixture

Triglycerides

(% w/w)

Diglycerides

(% w/w)

Monoglycerides

(% w/w)

0 100 — —

7 35 50 15

14 15 45 40

16 11 43 46

20 8 39 53

24 5 35 60

Source: Adapted from Ref. 1.

Figure 1 Reaction scheme of interesterification (glycerolosis) of triglycerides and

glycerol. R1, R2, and R3 are fatty acids.


glycerolysis contain from 40% to a maximum of 60% monoglycerides withthe balance being diglycerides and triglycerides. High-diglyceride-containingproducts can also be made by glycerolysis using 6 to 8 parts of glycerol to100 parts of fat. This ratio yields a blend containing 15% monoglycerides,50% diglycerides, and 35% triglycerides.

Concentrated monoglycerides are produced by a high-vacuum thinfilmmolecular distillation process yielding products containing typically 95%monoglycerides, 3–4% diglycerides, 0.5–1% free glycerol, and 0.5–1% freefatty acids.

Freshly distilled monoglycerides contain an equilibrium of 1-mono-glycerides and 2-monoglycerides. The ratio between the two isomers atdifferent temperatures is shown in Table 3. The composition varies consid-erably with temperature. The rate constant of the equilibrium reaction isvery low at room temperature and depends on fatty acid composition, crys-tal form, and traces of basic catalysts present (3). The content of 1-mono-glycerides in commercial distilled monoglycerides is usually from 90% toabout 95%.

Enzymatic synthesis of monoglycerides and other emulsifiers bas beenreported (7–9) as a method to produce stereospecific emulsifiers. However,isolation of the desired products from the reaction mixture is a problemfor large-scale production, and commercial production of emulsifiers bybiochemical methods is not commonly used.

Because the 1-monoglyceride isomer content can vary according to thetemperature history of the product, the most reliable method for the deter-mination of the monoglyceride content in commercial products is usinggas–liquid chromatography (GLC) combined with a derivatization of themonoglycerides with, for example, trimethyl silylether (4). Figure 2 shows agas chromatogram of distilled monoglycerides based on fully hydrogenatedtallow fat.

Monoglycerides are polymorphic like triglycerides and can existin different crystal forms depending on temperature conditions.Furthermore, as shown in Fig. 3, monoglycerides and diglycerides have ahigher melting point than the corresponding triglycerides. For saturated

Table 3 Equilibrium Reaction Mixtures of 1- and 2-Monoglycerides

Temperature

(�C)

1-Monoglycerides

(%)

2-Monoglycerides

(%)

20 95 5

80 91 9

200 82 18


monoglycerides of palmitic and stearic acid, the increase in meltingpoint compared to the corresponding triglycerides is 10–12�C. For unsatu-rated glycerides, like mono-olein and triolein, the difference is as high asabout 30�C.

The melting points of the polymorphs of pure monoglycerides withfatty acid chain length from C10 to C18 are shown in Fig. 4.

When cooled from melt, monoglycerides crystallize in a metastablea-form. On further cooling, a solid-state transition from an a-form to asub-a-form takes place at a lower temperature. When stored at ambienttemperature, transition to the stable b crystal form will take place. Thepolymorphism of monoglycerides thus follows the pattern of triglyceridesthat has been reviewed by Larsson (5).

All lipids form crystal structures with the hydrocarbon chainsarranged in layers with a parallel chain axis. In lipids with polar headgroups, such as surfactants, the molecule is always arranged in layers,with the polar head group in discrete layers separated by the fatty acidhydrocarbon chains, forming a lipid bilayer, which is an important feature

Figure 2 Gas–liquid chromatogram of trimethyl-silylether derivatives of distilled

monoglycerides. GL¼ glycerol, IS¼ internal standard, DIGL¼ diglycerol, SA¼

stearic acid, GMM¼ glycerol monomyristate, GMP¼ glycerol monopalmitate,

GMS¼ glycerol monostearate, GMA¼ glycerol monoarachidate, GMB¼ glycerol

monobehenate, GDP¼ glycerol dipalmitate, GPS¼ glycerol palmitate-stearate,

GDS¼ glycerol distearate.


of polar lipids. Figure 5 shows the molecular orientation in crystals of polarlipids with double-chain-length packing (DCL) and single-chain-lengthpacking mode (SCL) schematically, where the hydrocarbon chains of twoadjacent layers penetrate each other. The crystal forms of monoglyceridesand diglycerides as well as of distilled monoglycerides are the same as thosein triglycerides except that the b0-form is not usually found in commercialmonoglycerides.

X-ray diffraction patterns of commercial monoglycerides show astrong short spacing at 4.2 A for the a-form (hexagonal subcell), as seenin Table 4 showing crystallographic data and melting points of distilledmonoglycerides based on single fatty acids with minimum 90% purity orhydrogenated vegetable fats.

When the a-form is cooled, it transforms into a sub-a crystal form atapproximately 35�C; the sub-a-form is characterized by a strong spacingnear 4.3 A, and several spacings from 3.9 to 3.7 A, with medium intensity.

The stable, high-melting b-form is characterized by a strong shortspacing at 4.6 A, combined with several spacings in the region 3.9–3.6 Aor lower, with medium intensity.

Figure 3 Melting points of monoglycerides and diglycerides compared to

corresponding triglycerides with fatty acid chain length from C10 to C18. (From

Ref. 2.)


Figure 4 Melting points of crystal polymorphs of pure monoglycerides with chain

length from C10 to C18. C18 : 1t¼ 1-monoelaidin, C18:1c ¼ 1-mono-olein, C18:2c¼

1-monolinolein.

Figure 5 Schematic models of the molecular packing of polar lipids. DCL¼

double-chain-length structure, SCL¼ single-chain-length structure. Typical x-ray

diffraction long spacings of the lipid bilayers are shown.


Table 4 Melting Points and X-Ray Diffraction Data of Mixed Fatty Acid Distilled Monoglycerides

Monoglyceride

Melting pointsa

(�C)

Long spacings

(A)

Main short spacings

(A)

Sub-a a b a b a b

Monolaurin, 90% C12 16 45 61 — 37.3 4.15 4.57-4.32-4.00-3.84-3.71-2.44

Monomyristin, 90% C14 25 56 67 41.0 40.6 4.15 4.55-4.33-3.91-3.81-3.71-2.44

Monopalmitin, 90% C16 35 66 73 47.0 44.7 4.15 4.51-3.91-3.84-3.67-2.43

Mono-olein, 90% C18:1 — 30 34 — 48.5 — 4.60-4.38-4.31-4.04

Monobehenin, 90% C22 56 82 85 57.3 57.5 4.15 4.50-3.94-3.84-3.74-2.43

Saturated monoglyceridesb,1 37 71 75 54.0 51.4 4.13 4.55-3.94-3.86-3.78-2.43

(hydrogenated soybean oil)

Saturated monoglyceridesb,2 20 66 72 53.2 49.8 4.15 4.52-4.35-3.93-3.84-2.43

(from hydrogenated lard)

Saturated monoglyceridesb,3 16 68 72 51.6 47.0 4.15 4.55-4.33-3.89-2.43

(hydrogenated palm oil)

Unsaturated monoglyceridesb,4 (palm oil) 8 (48) 60 — 46.5 — 4.55-4.31-4.03-3.86

aDifferential scanning calorimetry (DSC) peak temperatures.bDanisco ingredients products: 1DIMODAN� HS, 2DIMODAN� HL, 3DIMODAN�HP, 4DIMODAN�P/M.


Long spacings of C16/18 saturated monoglycerides are in the order of50 A, corresponding to the DCL packing mode. Above the melting point,monoglycerides show a diffuse x-ray diffraction line in the long spacingregion corresponding to 30 A. This means that monoglycerides maintainan ordered structure in the melt due to hydrogen-bonding between thepolar groups.

Diglycerides are usually only present in monodiglycerides in minorquantities and are not considered to be the main components. However, itis possible to produce a high-diglyceride product containing 50–60%diglycerides by glycerolysis. Such products have found some applicationsdue to the specific crystallization properties of 1,2-diglycerides. Diglyceridesexist in two isomeric forms: 1,2-and 1,3-diglycerides (see Fig. 1) in a relativeratio of 40 : 60. The 1,2-diglycerides crystallize from melt in a metastablea-form, which transforms to a stable b0-form. They can, therefore, be usedto stabilize b0 crystals in fat blends, which otherwise show a tendency ofgrowing b crystals, giving rise to textural problems (e.g., in margarine orlow-calorie spreads).

B. Organic Acid Esters of Monoglycerides

The free hydroxyl groups in monoglycerides can be esterified with organicacids such as acetic, lactic, diacetyltartaric, citric, and succinic acids, andthus from more lipophilic or hydrophilic derivatives of monoglycerides. Theesters are normally produced by reacting monoglycerides with the organicacid directly or with its anhydride.

The crystalline behavior, melting point, and polar properties of theorganic acid esters are very different from that of the corresponding mono-glycerides. Acetic and lactic acid esters are low-polar, lipophilic emulsifiers,whereas diacetyl tartaric acid esters or citric acid esters are anionic, hydro-philic emulsifiers. Therefore, such organic acid esters have different func-tional properties than monoglycerides and are used in many newapplications in foods.

A common feature of the monoglyceride derivatives is that they are allmonomorphic and crystallize from melt in an a-crystal form. The meltingpoints and crystallographic data of organic acid esters are shown in Table 5.Some of the monoacyl esters crystallize with SCL packing mode, and theseemulsifiers have the strongest decrease in melting point compared to mono-glycerides (see Table 4).

The melting points shown in Table 5 are typical for monoglyceridederivatives based on blends of saturated palmitic and stearic acids. It shouldbe noted that among commercial products, some variation in melting points


Table 5 Melting Points and X-Ray Diffraction Data of Organic Acid Esters of Monoglycerides (C16/C18 ratio, 35 : 65)

Organic acid esters of monoglycerides

Melting point

(�C)

Long spacings

(A)

Chain packing

mode

Short spacings

(A)

Crystal

form

Acetic acid esters (monoacetylated) 39 32.9 SCL 4.10 aLactic acid esters (LACTEM) 42 39.5 SCL 4.13 aDiacetyl-tartaric acid esters (DATEM) 43 41.0 SCL 4.11 a1.2-Dipalmitin-DATEa (synthetic) — 54.6 DCL 4.21–3.76 —

Citric acid esters (CITREM) 60 60.3 DCL 4.11 a

aDiacetyl-tartaric acid ester of pure 1.2-dipalmitin.


could occur. This is due to variations in chemical composition and fatty acidprofile.

C. Acetic Acid Esters (ACETEM)

By reacting one or both of the free hydroxyl groups in distilled monoglycer-ides with acetic acid anhydrid and removing the surplus of free acetic acid bydistillation, a more lipophilic product is formed. Acetylation may be partialor complete depending on the ratio of the acetic acid anhydrid and mono-glyceride used. Normally, products with 50%, 70%, or 90% acetylation offree hydroxyl groups are used in foods, depending on application.

The melting point of ACETEM is about 20–30�C lower than that ofthe monoglycerides used in production and varies from 35�C to 40�Cdepending on the degree of acetylation and type of monoglyceride.Acetylation of unsaturated monoglycerides with iodine values higher than40 yields a product that is liquid at room temperature (melting pointapproximately 10�C).

The crystallization behavior of monoglycerides is changed consider-ably by acetylation, as ACETEM is monomorphic and stable in its a crystalform. Chemically pure 1-aceto-3-stearin shows the following polymorphictransitions when chilled from melt to below 0�C: sub-a1!3.5�C!sub-a2!12.5�C!a!45�C!b0!48.5�C!melt (6).

Acetylated, saturated monoglycerides of C16/C18 fatty acids form flex-ible films that can be stretched up to eight times their length before theybreak, and they are therefore used as coatings on fruits, nuts, and meatproducts (7). Due to its a-tending properties, ACETEM is also used infats (shortenings) in aerated foods such as cakes and toppings.

D. Lactic Acid Esters (LACTEM)

Production of LACTEM is normally based on a reaction between lactic acidand monoglycerides based on fully hydrogenated vegetable or animal fats.An alternative production method is esterification of glycerol, lactic acid,and fatty acids in a mole-to-mole ratio. Lactic acid esters may contain from15% to 35% esterified lactic acid distributed on a number of isomericcompounds. Some of the most typical components are shown in Fig. 6.

For LACTEM based on a C16–C18 chain length, saturated monoglyc-erides has a melting point of about 45�C. From melt, it crystallizes into astable a crystal form. X-ray diffraction of LACTEM often shows two longspacings at 38 A and about 55 A, indicating that some of the LACTEMcomponents crystallize in a SCL form (long spacing 38 A) and othercomponents crystallize in a DCL form (long spacing 55 A).


Generally speaking, LACTEMs are nonionic emulsifiers, soluble inoils and fats, and only slightly water dispersable. LACTEMs are oftenless surface active than their corresponding monoglycerides, depending onthe composition.

The LACTEMs are used mainly in foods such as cake shorteningsand fats for toppings or imitation creams and often in combinations withsaturated monoglycerides in order to make a product that is stable in the acrystalline form.

E. Diacetyl Tartaric Acid Esters (DATEM)

The anion-active and very hydrophilic DATEM products are produced byreacting diacetylated tartaric acid anhydride with monoglycerides. Thediacetylated tartaric acid anhydride is produced from refined, naturaltartaric acid from the wine industry by reaction with acetic acid anhydride.The chemistry of DATEM production is described in detail by Schuster and

Figure 6 Chemical formulas of (A) lactic acid ester of monoglycerides (LACTEM)

and (B) possible positional isomers of esters of lactic acid, palmitic acid, and glycerol.


Adams (8). The main components of DATEMs are referred to as DATEMI, II, III, and IV, and their schematic molecular formula is shown in Fig. 7.Depending on the type of monoglycerides used as raw material, a DATEMcan be crystallized in block, flake, or powder form or it can be semiliquid.

A DATEM of saturated, C16/C18 monoglycerides is stable in an acrystal form with a melting point usually about 45�C. X-ray-diffractionlong spacings are in agreement with a SCL packing of the hydrocarbonchains. It forms dispersions in water with low pH (2–3) due to its free-carboxyl group, and the solubility is increased if the pH value is adjustedto above 4–5 (see Section III). A DATEM is only partially soluble in oils

Figure 7 Chemical formulas of (A) diacetyl tartaric acid ester of 1-glycrol mono-

stearate (DATEM) and (B) positional isomers referred to as DATEM I, II, III,

and IV.


and fats. Compared to monoglycerides, a DATEM is very surface active andhas a number of applications in food emulsions. However, its main applica-tion is as a dough conditioner in yeast-raised bakery products.

According to the Food Chemical Codex in the United States, aDATEM may contain 17–20% by weight of esterified tartaric acid and14–17% esterified acetic acid. In Europe, the EU regulation prescribes awider variation in composition, allowing 10–40% esterified tartaric acid and8–32% acetic acid.

F. Citric Acid Esters (CITREM)

A CITREM is manufactured by esterification of monoglycerides with citricacid in an amount 12–20% by weight of the finished product, which is oftenneutralized partially, forming sodium salts of CITREM. CITREM manu-factured on the basis of saturated monoglycerides crystallize from melt in ana-like crystal form, and the melting point is between 55�C and 60�C.

All esters of dicarboxylic or tricarboxylic acids and monoglyceridesshow a high degree of long-range order in the melted state, a phenomenonwell known from soaps, called thermotropic mesomorphism. Low-anglex-ray diffraction of such materials shows one or several sharp lines in thelong spacing region at temperatures above the melting point. Citric acidesters exhibit this behavior, especially due to strong molecular interactionsbetween the polar groups in the melt.

A CITREM is an extremely hydrophilic, anionic emulsifier. It forms amilky dispersion in water and is only partially soluble in oils and fats. Itsmain application in foods is as antispattering agent in margarine or asemulsifier in meat or beverage emulsions. CITREM based on unsaturatedmonoglycerides is used to reduce the yield value and plastic viscosity of achocolate mix.

G. Succinic Acid Esters (SMG)

Succinic acid esters are produced by a reaction of succinic anhydride withmonoglycerides. A SMG is an anionic emulsifier, due to its free-carboxylgroup, as it appears from its chemical formula shown in Fig. 8.

A SMG based on saturated monoglycerides has a melting point of55–60�C. When cooled from melt, it goes through a thermotropic liquid-crystalline state before crystallizing in an a-form with SCL chain packing(x-ray long spacing 38 A). The a-form transforms to a stable b-form onstorage at room temperature. The b-form has a DCL chain packing asfound with monoglycerides.


Like a DATEM, a SMG is used mainly as a dough strengthener foryeast-raised baking products. However, its use in foods is limited to theUnited States, as EU regulations do not permit it.

H. Polyglycerol Esters of Fatty Acids (PGE)

Glycerol may be polymerised by dehydration and, by this reaction, aseries of polyglycerol compounds is obtained. They can be esterifiedwith edible fatty acids, usually palmitic stearic acid blends, forming poly-glycerol esters. A structure formula of triglycerol monostearate is shownin Fig. 9. Commercial PGE products may vary considerably in composi-tion, depending on the degree of polymerization and degree of esterifica-tion. According to EU regulations, the polyglycerol moiety should mainlybe diglycerol, triglycerol, and tetraglycerol with a maximum of 10% ofpolyglycerols equal to or higher than heptaglycerol. In the United States,however, the U.S. FDA (Food and Drug Administration) regulationpermits a polymerization degree up to decaglycerol. The compositionof PGE products is very complex, with a high number of positionalisomers.

A PGE with a low degree of polymerization (e.g., mainly triglyceroland tetraglycerol esters of C16/C18 fatty acids) melts at approximately 55�Cand crystallizes in a stable a-form. PGE is generally more hydrophilic thanmonoglycerides, but its dispersibility in water depends on polyol condensa-tion composition and degree of esterification. PGEs are used in a number offood emulsions, ranging from margarine and dessert products to cakes andother bakery products.

Figure 8 Chemical formula of succinic acid ester of glycerol monostearate.


Table 6 presents the crystallographic data and melting point of polyolfatty acid esters.

Purified diglycerol fatty acid esters can be manufactured by esterifica-tion of diglycerol with edible fatty acids. The monoacyl fatty acid esters are

Table 6 Melting Points and X-Ray Diffraction Data of Fatty Acid Esters of

Polyols and Lactic Acid (C16/C18 ratio, 50:50)

Product

Melting

point

(�C)

Long

spacings

(A)

Chain

packing

mode

Short

spacings

(A)

Crystal

form

Polyglycerol

monostearate 56 64.2 DCL 4.13 aPropylene glycol

monostearate 39 50.7 DCL 4.15–(3.99) aSorbitan monodistearate 54 54.5 DCL 4.11 aSorbitan monostearate — 33.8 SCL 4.11 aSorbitan tristearate 56 49.8 DCL 4.13 aSodium stearoyl lactylate 38 37.6 (49.7) SCL (DCL) 4.10 a

Figure 9 Chemical formula of a polyglycerol fatty acid ester [triglycerol

monostearate (PGE)].


concentrated by a molecular distillation process. The diglycerol mono fattyacid esters are more hydrophilic than corresponding monoglycerides (9).On the contrary, polyglycerol polyricinoleate (PGPR) is primarily used tostabilize low-fat water-in-oil (w/o) emulsions (spreads) or to reduce yieldvalue in chocolate mix.

I. Stearoyl-lactylates (SSL, CSL)

Esterification of stearic acid with lactic acid in the presence of sodium orcalcium hydroxides yields a mixture of stearoyl-lactylates (sodium or cal-cium salts), fatty acid salts, and free fatty acids. The main component ofstearoyl-lactylates is shown in Fig. 10. Esters of trilactic and polylactic acidsare also present in commercial products, making the composition quitecomplex.

The Na stearoyl-lactylates (SSL) are normally present in an a-formwith a melting point of about 45�C. This form is obtained by spray-coolingthe melt into a powdered or beaded product. The a-form has a single shortspacing at 4.1 A and long spacings at about 38 A, showing the SCL packingmode with penetrating hydrocarbon chains.

The SSL is water dispersible at neutral pH. At a pH value below 4–5,solubility is limited due to the content of 15–20% free fatty acids in SSL.The CSL is less water dispersible, but more oil soluble than SSL. Both SSLand CSL are used mainly in the baking industry as dough strengtheners.SSL is also used in many emulsions such as imitation creams, coffee whit-eners, filling creams, and icings.

J. Propylene Glycol Esters of Fatty Acids (PGMS)

Propylene glycol esters can be manufactured by esterifying propylene glycolwith edible fatty acids under alkaline catalysis at about 200�C under

Figure 10 Chemical formula for stearoyl-lactoyl lactic acid sodium or calcium salt

(SSL, CSL).


vacuum. After removal of excess propylene glycol, the reaction blendcontains approximately 55% propylene glycol monoester and 45% diester.Structural formula is shown in Fig. 11. An alternative method is interester-ification of triglycerides and propylene glycol, yielding a reaction mixturecontaining the propylene glycol monoesters and diesters, together with10–15% monoglycerides and a minor amount of diglycerides as well astriglycerides. The propylene glycol monoester can be concentrated by amolecular distillation process as described for monoglycerides. DistilledPGMS contains a minimum of 90% monoesters and is normally based onsaturated C16/C18 chain length fatty acids.

Commercial propylene glycol monostearate is stable in an a-like form,as shown in Table 6. However, pure 1-propylene glycol monostearate existsin four different crystal forms (10).

The PGMS are only slightly water dispersible, but completely solublein oils and fats. Therefore, it is used only in combinations with fats of as ana-tending emulsifier in combination with monoglycerides or other emulsi-fiers. PGMS is mainly used in cake shortenings and fats for whippableemulsions, toppings, and so forth.

K. Sorbitan Esters of Fatty Acids

1. Sorbitan Esters of Fatty Acids (SMS, STS)

Sorbitan is derived from sorbitol by dehydration and then esterified withfatty acids and described for PGMS, and depending on the amount of fattyacids used for the esterification, sorbitan monoesters (SMS) or triesters(STS) are produced. The chemical formulas of SMS and STS are shownin Fig. 12.

All sorbitan esters are stable in the a crystal form, but their x-ray longspacings indicate a different molecular packing depending on the degree ofesterification, as shown in Table 6. Commercial sorbitan monodistearate orsorbitan tristearate crystallize with a DCL chain packing mode, whereas aconcentrated sorbitan monostearate with a mono-acyl content of 80% has along spacing of about 34 A, corresponding to SCL chain packing.

Figure 11 Chemical formula of propylene glycol monostearate (PGMS).


Sorbitan tristearate (STS) is very lipophilic and its main application isas a crystal modifier in fat-based foods (margarine, spreads, chocolate pro-ducts). STSs are used to control crystallization of fats, where they stabilizethe b0 crystal form, preventing formation of the higher melting b crystalform. The b crystals tends to grow very large and cause a grainy texturein margarine or spreads. In chocolate products, similar fat crystal transi-tions are the reason for the development of a fault referred to as ‘‘bloom,’’which appears as grayish spots on the surface.

Sorbitan monostearate (SMS) is dispersible in warm water and solublein oils and fats, whereas STS is soluble in oils and fats only.

Sorbitan monostearate is used in many food products, primarilyemulsions, and often in combination with the ethoxylated sorbitan esters(polysorbates).

2. Polyoxyethylene Sorbitan Esters

The hydrophilic properties of sorbitan esters can be strongly increased byethoxylating the free-hydroxyl groups, yielding an ethoxylated sorbitan

Figure 12 Chemical formula of sorbitan fatty acid esters: (A) sorbitan

monostearate [6-stearoyl-1,4-anhydro-D-glucitol (SMS)] and (B) sorbitan tristearate

(STS).


ester. Such products are referred to as polysorbates and are plastic or liquidproducts, completely soluble in water, where they form micelles. Their pro-duction and properties are described in detail by Benson (11).

Polysorbates are highly surface active and are used in many technicalemulsions, but their application in foods is somewhat limited due to theirlow acceptable daily intake (ADI) value.

L. Sucrose Esters of Fatty Acids

Sucrose esters have been known for more than 30 years as esterificationproducts of fatty acids and sucrose and other sugars, or as mixtures ofpartial glycerides and sucrose esters made by trans-esterification of sugarsand glycerides. A great number of components from monoesters to octa-esters can exist in sucrose esters, depending on the degree of esterification.

There are several ways of producing sucrose esters (11,12) using dif-ferent solvents or without solvent systems. The production may includesolvent extraction for separation of esters with different hydrophilic–lipophilic characteristics. Sucrose monostearate is highly water soluble andcan be used as an o/w emulsifier. Sucrose esters are generally accepted inJapan, but in other countries, the use of sucrose esters is limited in foodsystems.

M. Lecithin

Plant-seed lecithins are used in many foods, and lecithins from egg yolk havebeen used in margarine for many decades. The most common commerciallecithins today stem from soybean oil. The trade name lecithin is used com-monly, although the product consists of a blend of different phospholipids.

Soybean lecithin is normally marketed as a solution of phosphatides insoybean oil containing between 60% and 70% phosphatides. Soybeanlecithin can be fractionated in different ways (e.g., with acetone to removemost of the triglycerides present.

The composition of acetone-precipitated soybean phospholipids isapproximately 41% phosphatidyl choline, 34% phosphatidyl ethanolamine,19% phosphatidyl inositol, and 6% other phospholipids.

The isolated phosphatides may be further fractionated by ethyl alcoholor other short-chain alcohols, giving an alcohol-insoluble part consistingmainly of phosphatidyl choline. Lecithin may be hydrolyzed by enzymetreatment or reacted with hydrogen peroxide to make it more hydrophilic.A synthetic phosphatide for use in the chocolate industry, ‘‘Emulsifier YN,’’is manufactured by the reaction of a diglyceride with phosphorus pentaox-ide and neutralizing the reaction product partly with ammonia.


A comprehensive review of industrially produced lecithins and theirapplications are given by Szuhaj and List (13).

N. Structural Parameters of Emulsifiers

Long spacings obtained by x-ray diffraction analysis of homolog series ofemulsifiers based on single fatty acids with varying chain lengths can giveinformation about the molecular packing conditions in the crystalline state(14). Such data have been reported for monoglycerides, DATEMs and sor-bitan esters (15). The size of the polar group in the crystalline lattice is ofspecial interest, and for distilled monoglycerides, a value of 5.3 A was found.This is quite similar to the polar group size of pure 1-monopalmitin (5.5 A).The polar head group of DATEMs was found to be 9.3 A, which is in thesame range as pure 1,2-dimyristyl phosphatidyl choline (10.4 A), as reportedby Small (16). Sorbitan esters were found to have a polar head group size of5.4 A, quite similar to the dimension of glycerol in monoglycerides.Furthermore, the x-ray data indicate that in sorbitan tristerate crystals,the sorbitan group is oriented in the middle of the lipid bilayer, withone acyl group extended to one side and the two other acyl groups tothe other. This indicates a similar structure to triacylglycerides and mayexplain why sorbitan tristearate is an effective crystal modifier in fat-basedfoods (17).

III. LYOTROPIC MESOMORPHISM

Humans have used aqueous liquid-crystalline phases in soap preparationsfor more than 2000 years. However, liquid crystals were first mentioned in1904 by Lehman (18), and the structure of such polar lipid–water systemshas been known only for about 50 years. Until about 1960, it was believedthat both the so-called ‘‘neat’’ and ‘‘middle’’ soap phases had a lamellar-type structure. One of the first important attempts to clarify liquid crystal-line structures was done by Luzzati and his co-workers in 1960 (19) whenthey demonstrated how to analyze lipid–water structures by low-angle x-raydiffraction. After that, numerous articles on mesomorphic lipid–waterphases, their structure, phase behavior, and occurrence in biological systemshave been published. Especially, soap systems and phospholipid–water sys-tems have been studied extensively. A comprehensive review of surfactant–water liquid crystals has been given by Tiddy (20), and the phase behavior ofphospholipids has been reviewed by Hauser (21); finally, the contribution bySmall (16) gives the most detailed review on lipid–water interactions. Basedon this background, only a brief description of lyotropic mesophases formed


by polar lipids will be given here, together with a survey of the mesomorphicproperties of food emulsifiers.

The molecular orientations of polar lipids in their crystalline state anda lamellar liquid-crystalline phase are shown in Fig. 13a. In their crystallinestate, the lipids are orientated in bilayers with the polar groups ‘‘head tohead,’’ separated by layers of solid hydrocarbon chains usually packed in aDCL mode. Lipids with large polar groups may form crystals with penetra-tion of hydrocarbon chains in a SCL mode as described earlier. When wateris present and the temperature is above the Krafft point, Tc, water willpenetrate through the polar region of the crystals. At the same time, transi-tion from solid state to liquid state takes place in the hydrocarbonchain region, resulting in the formation of a liquid-crystalline phase. Suchbehavior is called lyotropic mesomorphism.

When a polar lipid in a stable crystalline form (e.g., b crystals) is mixedwith water and heated to form a liquid-crystalline phase, less energy isrequired per mole of lipid than heating it in the absence of water to forma melt. This is demonstrated in Table 7, showing the transition enthalpy(�H) for b crystal transition to melt and b crystalþwater transition to aliquid crystal of pure 1-monopalmitin (1-GMP) and of distilled, saturatedmonoglycerides (C16 : C18 ratio¼ 35 : 65). The experimental value of �H aremeasured by differential scanning calorimetry (DSC).

It appears from Table 7 that the enthalpy values for the 1-monopal-mitin transitions are higher than the corresponding values for distilled

Figure 13 Structure models showing (a) the orientation of emulsifier molecules in

the DCL structure in the crystalline state, (b) the formation of a lamellar, liquid-

crystalline phase above the Krafft point (Tc) in water, and (c) the formation of an

a-gel phase at a temperature (T) below Tc. The structure parameters d, da, and dwcan be measured by x-ray diffraction analysis.


monoglycerides with a mixed fatty acid composition. This is probably due toa higher entropy in the distilled monoglycerides than in the pure 1-mono-palmitin.

The latent heat of crystallization dissipated when the a-gel phasetransforms to b crystalsþwater is a measure of the instability of the a-gelphase.

A. Lamellar Phase

The structure of the lamellar phase is shown in Fig. 13b and consists ofbimolecular lipid layers separated by layers of water between the polargroups. The hydrocarbon chains are in a liquidlike state with a higher degreeof motional freedom than in the corresponding anhydrous melt. Themolecules in the lipid bilayer of the liquid-crystalline phases undergo sev-eral kinds of motion such as lateral diffusion, rotation, and the so-calledflip-flop, which is a change in position from one side of the bilayer toanother.

Figure 13b shows the structure parameters d, da, and dw of the lamellarphase, which can be measured by low-angle x-ray diffraction methods.

The swelling of the lamellar phase can vary from 10 to 20 A for thewater layer thickness in pure 1-monoglyceride–water systems (22) to water

Table 7 Transition Energies (�H) of Monoglycerides Compared to

Monoglyceride–Water Systems

Transition scheme

Pure 1-monopalmitin Distilled monoglycerides

Temp.

(�C)

�H

(J/g)

Temp.

(�C)

�H

(J/g)

Bulk lipid 75.9 210 70.0 172

b Crystals!melt

Melt!a crystal 65.2 �113a 65.0 �105

a!Sub-a 37.7 �43 19.0 �17

Sub-a!b crystalb — �54 — �50

Monoglyceridesþwater systems 60.3 203 55.0 133

b Crystalsþwater

! lamellar lipid crystal

Lamellar liquid crystal

!a-gel phase 47.7 �84 55.0 �63

a-Gel!b crystalsþwaterb — �119 — �70

aNegative �H values¼ exothermic transition.bCalculated values.


layers of several hundred angstroms in commercial monoglyceride–watersystems containing small amounts of ionised free fatty acids (23).

B. Phase Behavior of Monoglyceride–Water Systems

The mesomorphic behavior of emulsifiers in water depends on a number offactors:

1. Wedge shape of the surfactant molecule (i.e., the relative size ofthe polar head group to the length and degree of unsaturation ofthe hydrocarbon chain)

2. Temperature and ratio of surfactant to water3. Degree of ionisation of ionic surfactants (pH). Ion concentration in

the aqueous phase

In general, a blend of surface-active lipids (polar and nonpolar) maybehave like one component, with the average polar characteristics like thatof the blend. The phase behavior of surfactants in water can, therefore, becontrolled by adjusting the average polarity by the addition of polar surfac-tants to less polar ones and vise versa.

Distilled monoglycerides, with approximately 95% 1-monoglyceridecontent, form similar mesomorphic phases in water as do pure 1-mono-glycerides, but show different swelling properties. When saturated, distilledmonoglycerides in their b-crystalline form are mixed with water and thesuspension is heated to a certain temperature (Krafft point); when themelting point of the hydrocarbon chain bilayer is reached, then water willpenetrate through the planes of the polar groups and a lamellar mesophaseis formed with water layers alternating with lipid bilayers (see Fig. 13b).

The degree of swelling (i.e., the thickness of the water layer) may varyconsiderably, depending on a number of factors. The hydration force orosmotic repulsion will induce swelling of monoglycerides to a water layerthickness of about 16 A. At this interplanar distance of bilayers, there is abalance of long-range van der Waals forces that keep the bilayers togetherand the hydration force. With pure monoglycerides, swelling is thereforelimited to a level corresponding to a water layer thickness of 16–20 A (22).

Electrostatic repulsions may further increase the swelling process, sothe water layer thickness may increase to several hundred angstroms. Thiscan be obtained by the incorporation of ionic polar lipids in the lipidbilayers. Swelling may then increase continuously in proportion to thewater content of the system (23).

Binary phase diagrams of distilled, saturated, and unsaturated mono-glycerides, pure glycerol monoelaidate, and pure glycerol mono-oleate areshown in Fig. 14. The distilled, saturated monoglycerides are based on fully


hydrogenated animal fat (lard or tallow), a commercial product (DimodanPM, from Danisco, Denmark). This product contains mainly C16/C18 fattyacids and will be referred to as DGMS.

Studies of DGMS–water systems by means of a temperature-programmed x-ray diffraction method have shown that the phase transitionfrom crystalþwater to the lamellar phase takes place via formation of ana-gel phase. The a-gel phase exists in a temperature region of about 5�Cbelow the region of the lamellar phase or dispersion state at high watercontents. Such behavior is not found with pure monoglycerides. Thephase diagram of pure 1-monopalmitin–water (22) shows a direct transition

Figure 14 Binary phase diagrams of (a) distilled, saturated monoglycerides with

fatty acid chain length C16/C18 (35 : 65, DIMODAN HA, Danisco, Denmark), (b)

distilled, unsaturated monoglycerides based on sunflower oil (DIMODAN U,

Danisco, Demark), (c) pure 1-monoelaidin, and (d) pure 1-mono-olein. Abbrevia-

tions: Tc¼ Krafft temperature; Tg¼ transaction temperature for b crystalþwater to

a-gel; L¼ lamellar; G¼ gyroidal structure; D¼ diamond structure. (Modified from

Refs. 24 and 25.)


from the b-crystalline state to a lamellar phase. The formation of the a-gelphase in DGMS–water systems may be due to the mixture of different fattyacid chain lengths and the presence of a minor content of free fatty acids andfree glycerol (usually below 0.5%). Apart from the formation of an a-gelphase, the DGMS–water phase diagram is quite similar to that of pure1-monopalmitin.

The distilled, unsaturated monoglycerides referred to in Fig. 14b aremade from sunflower oil. The fatty acid composition is 22% oleic acid,64% linoleic acid, and 13% saturated C16 and C18 fatty acids. The pure1-monoelaidate and 1-mono-oleate were made synthetically from 99% pureelaidic and oleic acid.

The phase behavior of distilled unsaturated monoglycerides is domi-nated by the viscous, isotropic cubic phase, which is formed at room tem-perature at water contents above 20%. The cubic phase is of the gyroidaltype and can accommodate up to approximately 35% water (26). At higherwater contents, a two-phase system of the cubic phaseþwater exists. Athigher temperatures, from 55�C to 65�C, the cubic phase transforms to ahexagonal II phase; above 90�C, a fluid isotropic L2 phase exists with excessfree water at water contents above 30%.

Monoglycerides of elaidic and oleic acids are commonly present incommercial monoglycerides based on partially hydrogenated vegetableoils. Therefore, their phase diagrams in water are shown in Figs. 14c and14d. It can be seen from Fig. 14c that 1-monoelaidin forms a lamellar phaseat 33�C and has a very narrow temperature range for the dispersion state athigh water contents. The dominating mesophase in the 1-mono-olein–watersystem is the cubic gyroidal phase, which exists at a temperature range frombelow 20�C to 90�C at water contents above 20%. Two types of cubic phaseare shown: the gyroidal (G) phase, which exists at water contents from 10%to about 35%, and the diamond (D) phase, which exists at higher watercontents than 35% (24).

The phase behavior in water of industrial distilled monoglyceridesfrom different manufacturers may vary slightly due to small changes incomposition. It is important that the product be completely free oftriglycerides, because minor quantities thereof will change the phasebehavior markedly. A minor content of diglycerides (3–5%) is normallyalways present in distilled products and does not affect the formation ofmesophases.

Saturated, distilled monoglycerides are often used in food pro-cessing as an aqueous mixture to obtain good distribution of the mono-glycerides in the product. This happens particularly in cases wheremonoglycerides cannot be added to the product via a suitable carrier,such as fats or oils.


When DGMS–water systems were first studied (25), it was discoveredthat a neutral pH and a low electrolyte concentration in the water was ofgreat importance for the formation and stability of diluted dispersions ofDGMS in water. When very dilute dispersions containing 5–10% DGMSwere buffered to pH 7, a clear homogeneous dispersion was obtained, andwhen cooled down, a stable gel could be formed. If the pH value was as lowas 5–6, the dispersion was not clear, but milky, and the gel transformedquickly into a coagel (b crystalsþwater). In this connection, the presence ofsmall quantities of free fatty acids plays an important role. It was laterdemonstrated (23) that the neutralization of the free fatty acids in DGMSincreases in the swelling capacity of the lamellar phase to an optimum level.

Figure 15 shows x-ray spacings of lamellar phase of distilled mono-glycerides made from fully hydrogenated fat (C16/C18: 30/65) in water at60�C (23). Curve a represents system with neutralized free fatty acids (pH7.0), and curve b represents systems with distilled water (pH 4.0), where theamount of free fatty acids (0.5%) in monoglycerides is not neutralized.There is a great difference in the swelling properties of the two system.

Figure 15 Swelling behavior of distilled, saturated monoglycerides (DIMODAN

HA) in water at 60�C (a) with neutralization of 0.2% free fatty acids and (b) without

neutralization. The interplanar spacing corresponds to the structure parameter d in

Fig. 13b.


In the neutralized system (a), and d spacings increase in a linear relationshipto the water content, showing continuous swelling behavior. Becausedw¼ d� da and da¼ 38 A, it means that at a water content of 75%, thewater layer thickness (dw) is about 120 A. The swelling of such neutralizedmonoglyceride–water systems continues with higher water contents, but nox-ray data are available for such diluted systems.

Contrary to the swelling behavior in neutralized water systems (pH7.0), monoglycerides show a limited degree of swelling in distilled water witha pH below 4.0. Curve b in Fig. 15 shows a maximum d value of 54 A, at awater content of 30%, corresponding to a water layer thickness of 16 A. Athigher water contents, no further increase in the d value is found, showingthe same limited swelling behavior as found in pure 1-monoglycerides. Theeffect of introducing ionised amphiphilic molecules in the lipid bilayer is anincreased swelling due to electric repulsion between charged groups in oppo-site bilayers. This repulsion effect will be reduced if the ion concentration ofthe water phase (e.g., Naþ Cl�) is increased. At concentrations above 0.3%sodium chloride in water, the effect of adding ionic active emulsifiers to thelamellar phase is inhibited (23).

C. Dispersion State and Crystalline Hydrates

All saturated monoglycerides form a dispersion as defined by Larsson (22)within the temperature region of the lamellar phase, when the water contentof the system is higher than that corresponding to the lamellar phase. It isassumed that the dispersion consists of lamellar aggregates of monoglycer-ides in equilibrium with water.

The thickness of the water layer in the dispersion aggregates may varyconsiderably depending on the conditions with regard to pH, ion concen-tration, and purity of the monoglycerides. The dispersion state does notexist at temperatures above the transition point of lamellar phase to cubicphase. When the temperature of the dispersion is increased above that point,a two-phase mixture consisting of cubic phaseþwater is formed. Therefore,in diluted aqueous dispersions of monoglycerides, the temperature shouldalways be carefully controlled in order to avoid formation of the cubicphase. Crystallization from a dispersion of monocaprin can be seen inFig. 16, where a fragment of a b crystal is visible in the lower left corner.It can also be seen that a number of dispersion aggregates are in contactwith the crystal surface. On the basis of similar observations, a specific modeof crystallization has been suggested which will lead to the formation ofb crystals with unusual properties (27). It is proposed that the bimolecularlipid leaflets unfold from the dispersion aggregate into the surface of thecrystal, which will then grow with discrete layers in a direction perpendicular


to the surface plane. This direction of growth can, in fact, be seen directlyunder the microscope in a dispersion, as shown in Fig. 16. The result of sucha crystallization process will be surface consisting of the polar groups of themonoglycerides. Normally, the crystal surface consists of the lipophilicmethyl end groups when crystallization takes place from the melt or fromsolvents. The polar surface formed when crystallized from an aqueous phaseis responsible for a great water-binding capability. For example, 10 parts ofmonoglycerides by weight can absorb 90 parts of water; such a mixture hasan ointment like consistency. Industrial products of 20–25% distilled, satu-rated monoglycerides in the form of b crystals suspended in water (so-calledhydrates) are commercially available and are used to a great extent in thebaking industry as crumb-softening or antistaling agents in bread.

The rate of crystallization initiated by cooling of the dispersion to atemperature below Tc depends on various factors such as the chain length ofthe monoglycerides and pH. Monoglycerides with short chain length, likemonocaprin or monolaurin, recrystallize relatively quickly, compared topalmitic and stearic fatty acid monoglycerides. Pure 1-monoglycerideswith chain lengths shorter than monomyristin transform directly fromthe lamellar phase to a b crystalþwater mixture (22). At a low pH, the

Figure 16 Photomicrograph of a dispersion of monocaprin in water, showing

sperical dispersion aggregates with internal lamellar structure. A fraction of a

b crystal of monocaprin is seen in the lower left corner. Bar: 30 mm, polarized light.


addition of salts and mechanical agitation will promote the formation ofb crystals, and this is used in the production of monoglyceride hydrates on acommercial scale.

D. Mesomorphic Phases of Other Food Emulsifiers

In general, the phase diagrams of the commercial food emulsifiers describedin this sub-section will show the typical mesomorphic phase behavior ofthese products without going into detail about multiphase regions thatmay exist.

1. Polyglycerol Esters

The composition of commercial polyglycerol esters is very complex due tovariation in both the degree of polymerization and degree of esterification.This results in products containing many different components. Therefore,when polyglycerol esters of C16/C18 fatty acids are mixed with water abovetheir melting point, they form reversed hexagonal phases. The hexagonalphase swells up to a content of 40% water, and at higher water concentra-tions a two-phase mix of emulsifier and water is obtained. The reversedhexagonal phase is common, with polar lipids containing a mix of polarand nonpolar components. Monodiglycerides with a mono ester contentbelow approximately 60% form hexagonal phases in water (22).

The dispersibility in water of such lipophilic emulsifiers can beimproved by adding high-polar coemulsifiers (e.g., sodium stearate).Commercial polyglycerol esters may contain up to 6% sodium stearate;this solubilizes the hexagonal phase, making it possible to produce aqueousdispersions or gels to be used as aerating agents in food products. Theemulsifier is mixed with water at a temperature above its melting point(Krafft point) until total swelling is obtained; then, the mix is cooled toambient temperature, forming a plastic gel. Such gels are used in thebaking industry to facilitate aeration of cakes of various types. Thephase behavior of triglycerol and octaglycerol esters have been describedelsewhere (28).

Distilled diglycerol monoacyl fatty acid esters have a unique phasebehavior in water. All diglycerol esters with chain lengths from C12 to C18

form lamellar phases above their Krafft point, and in contrast to the phasebehavior of monoglyceride–water systems, no other liquid-crystalline phaseshave been observed. Even the unsaturated diglycerol mono-olein forms alamellar phase only in the temperature region examined. The Krafft point ofdistilled diglycerol esters with fatty acid chain lengths from C12 and C18 isshown in Fig. 17.


The swelling capacity of diglycerol esters in water is limited, and abovea lipid : water ratio of 60 : 40, a stable dispersion of multilamellar vesicles isformed. The water layer thickness in the vesicles is 22–30 A (15), corre-sponding to the swelling behavior of diacyl phosphatidyl choline (16).The addition of anionic coemulsifiers (e.g., sodium stearate) increasesthe swelling properties of diglycerol esters in a fashion similar to that ofmonoglycerides containing fatty acid sodium salts.

Distilled diglycerol esters of stearic or oleic acid are capable of formingliquid-crystalline films around oil droplets in o/w emulsions and can be usedto stabilize protein-free emulsions (9).

2. Organic Acid Esters

The esters of dicarboxylic acids such as diacetyl tartaric acid or succinic acidand saturated monoglycerides (DATEM, SMG) contain a free-carboxylgroup. Neutralization of the free-carboxyl group, by forming a sodiumsalt of DATEM or SMG, increases its swelling ability in water. X-raydata of the lamellar phase formed by DATEMs at 60�C show asimilar swelling behavior to that of the neutralized DGMS–water systems.A schematic phase diagram of neutralized DATEMs in water is shown inFig. 18.

Figure 17 Krafft point lines of distilled, diglycerol fatty acid esters with fatty acid

chain lengths from C12 to C18. The formation of lamellar liquid-crystalline phases

and dispersion of vesicles above the Krafft point of the emulsifiers as a function of

the water content are indicated.


A dispersion of 5% DATEMs in water has a pH value of 1–2, and atthis low pH, very little, if any, swelling of DATEMs in water takes place.

An increase in pH, by adding sodium hydroxide to the water until apH of 4.6 is reached, will give total swelling in water. The DATEMs basedon saturated monoglyceride form lamellar phases above a temperature of45�C. The DATEMs based on an unsaturated monoglyceride, such asmono-olein, form a lamellar phase at 20�C, which is stable up to above80�C. Once the DATEM has swelled in water, it remains in its liquid,crystalline state, even when cooled below the temperature where the lamellarphase was formed. The hydrocarbon chains have little tendency to crystal-lize and form a gel. The DATEM esters are used mainly in the bakeryindustry in yeast-raised doughs, where their interactions with flour compo-nents improve the volume and texture of the finished product. The anionicemulsifiers (e.g., DATEM, SSL) are used in various o/w-type emulsions inorder to increase stability towards coalescence, due to their interaction withinterfacial bound proteins.

3. Stearoyl Lactylates

Sodium stearoyl lactylates (SSLs) are not neutralized completely. They havean acid value from 60 to 80, and the pH of an aqueous dispersion of SSL istherefore about 4–5. The phase behavior of SSL in water is strongly depen-dent on its degree of neutralization. At low pH conditions, a hexagonal IImesophase is formed at temperatures above 45�C, and this phase may

Figure 18 Schematic phase diagram of diacetyl tartaric acid ester of saturated

monoglycerides (DATEMs) in water, pH5.


contain 20–40% water as shown in Fig. 19a. If more water is present, theexcess water will separate from the mesophase. X-ray data concerning sucha hexagonal II mesophase of SSL–water systems at 60�C and pH 4.5 haveshown that the diameter dw of the water cylinders increases from 24 A at20% water to 40 A at 42% water and that the specific surface in contact withwater (S) increases from 25 A2 to 43 A3 with increasing water content from20% to 42%. The structural parameters S, da, and dw are very similar to

Figure 19 Schematic phase diagram of sodium stearoyl lactylate (SSL) in water

(a) with partial neutralization to pH 5.0 and (b) after neutralization to pH 7.0.

The figure shows an increased swelling into a lamellar phase or a dispersion of

vesicles at pH 7.0.


data reported by Reiss Husson (29) for egg phosphatidyl ethanolamine,which also forms a hexagonal II mesophase in water. When SSL is neutra-lized, its phase behavior is changed, and a lamellar phase is formed, insteadof the hexagonal II phase, as shown in the phase diagram in Fig. 19b. Thelamellar phase of the neutralized SSL can be diluted with water to disper-sion. The gel formed from the lamellar phase on cooling is very stable, andSSL may be used together with other emulsifiers (i.e., monoglycerides inorder to stabilize gel phase, which are used in food processing. Due to itshydrophilic nature, SSL is used in many o/w emulsions. Both SSL and thecorresponding calcium salt (CSL) are effective starch-complexing agentsand are used mainly in the bakery and starch industries.

4. Lecithin

Lecithin is a natural emulsifier that has been used in foods for centuries. Thephase behavior of lecithins in water was been reviewed in detail (16,21), soonly a brief description of their phase behavior will be given here. Purephosphatidyl choline forms a lamellar mesophase in water at temperaturesfrom ambient to 100�C. The swelling of the lamellar phase is limited to a40% water content, corresponding to a water layer thickness of about 30 A.At higher concentrations of water, a dispersion of spherical aggregates withinternal lamellar structure in the excess water (liposomes) is formed.Phosphatidyl ethanolamine (cephalin) is reported to form both lamellarand hexagonal II phases in water (29), whereas egg yolk lysolecithin exhibitsa different lyotropic behavior than that of the other egg yolk phosphatides.Due to its more hydrophilic character, it forms hexagonal I phases in waterin concentrations of up to about 55% water at temperature above 37�C.Commercial lecithins from soybean or rapeseed are mixtures of phospholi-pids with different phase behaviors in water. Mixtures of phosphatidylcho-lin and phosphatidylinositol form only lamellar phases, whereas the morelipophilic phosphatidylethanolamine forms reversed hexagonal phases inwater. Consequently, the phase behavior of oil seed lecithins may varywith their composition.

5. Polysorbates

When fat derivatives such as monoglycerides or sorbitan esters are reactedwith ethylene oxide, the resulting products become more hydrophilic incharacter, and their ability to form micelles in water is greatly increased.Polyoxyethylene (20) sorbitan monoleate (POESMO) forms a mesomorphicphase in water of the hexagonal I type with lipid cylinders surrounded bywater, at temperatures up to 30�C, and with water contents from 30% upto 60%. POESM-stearate also produces this phase with the same amount of


water, but at temperatures about 20�C higher (i.e., from 38�C to 50�C). Athigher temperatures, an isotropic solution is formed. X-ray data of aqueoussystems of these products (23) have shown that the specific surface S permolecule in contact with water is 145 A2, which is four times higher than thespecific surface of monoglycerides in the lamellar phase, as shown in Table 8.The hexagonal I mesophase (middle) can solubilize triglycerides in smallquantities. On further addition of triglycerides, a lamellar phase is formed.

Sorbitan monostearate has, in general, a similar phase behavior inwater to that of glycerol monostearate, but it has not been studied indetail. It forms a lamellar phase, and when excess water is added, aliquid-crystalline dispersion is made. Sorbitan tristearate, however, doesnot form any mesomorphic phases in water. The use of polysorbates andsorbitan esters in emulsions of various types is reviewed by Benson (11).

E. The a-Crystalline Gel State

The term ‘‘gel’’ used in this context originates from the soap industry manydecades ago. It was used then to describe certain soap–water systems with aclear gellike appearance. The term should not be confused with gels formedby other substances, such as starch, or hydrocolloids such as gelatin or pectin.

The gel is a crystalline state of the emulsifier with layers of waterbetween the polar groups, as shown in Fig. 13c. The structure of the gel issimilar to the lamellar phase. The difference is that the hydrocarbon chainsare no longer in a liquidlike state, but are solid and orientated parallel toeach other in an a-crystalline mode of packing. Consequently, the gel ischaracterized by a single strong x-ray diffraction line at 4.15 A in theshort spacing region, and several long spacings with an interrelation of1 : 1/2 : 1/3 : 1/4, indicating a lamellar system. The angle of tilt of the hydro-carbon chains toward, the water layer in a gel is 54�, and the lateral packingof the chains can be described by a hexagonal subcell, where the chains haverotational freedom (22). The gel is a metastable state, and the water layerbetween the polar groups will be reduced in time. When all the water isexpelled, a suspension of crystalsþwater (coagel) is formed. The so-calledhydrate form of monoglycerides is a coagel. Polymorphic emulsifiers such asmonoglycerides will transform into the most stable crystal form, the b-form,because an anhydrous a-form does not exist in the presence of water.Emulsifiers that are nonpolymorphic and stable in the a-crystalline formmay form gels with water that are very stable. This is the case with gels ofsodium stearoyl lactylates, tetraglycerol monostearates, SMG, DATEMs,and polysorbates. The stability of monoglyceride gels depends on theircomposition, because monoglycerides with fatty acid chain lengths offrom C16 to C20 are more stable in the gel form than monoglycerides with


Table 8 X-ray Data of Lamellar Mesophases and Corresponding Gel Phases of Monoglyceride-in-Water Systems

Compared to Other Emulsifier-in-Water Mesophases

Emulsifier-in-water system

X-ray

spacings d

(A)

Lipid bilayer

thickness da(A)

Water layer

thickness dw(A)

Specific surface in

contact with water S

(A2)

Monoglycerides

Lamellar phase, 40% water, 60�C 63.8 38.1 25.7 32.0

Gel phase, 40% water, 25�C 93.9 54.9 39.0 22.0

Lamellar phase, 75% water, 60�C 154 38.0 116 32.0

Gel phase, 75% water, 25�C 230 55.0 175 22.0

Polysorbate-60

Hexagonal I phase, 60% water, 45�C 90.0 59.8 30.2 145

Na-stearoyl lactylate

Hexagonal II phase, 41% water, 60�C 59.5 19.2 40.3 42.6


shorter chain lengths. The swelling capacity and stability of industrial dis-tilled monoglyceride gels are strongly influenced by the presence of ionicactive emulsifiers and salt concentration in water (23,30).

When cooled the neutralized DGMS–water mixtures shown in Fig. 15form gels with water contents up to about 75%, and the x-ray spacingsincrease in direct proportion to the water content up to about 230 A at75% water (30). No spacing were ever found in the region between 64 Aand 90 A, corresponding to water contents between 20% and 40% in theneutralized system. A similar phenomenon has been described for soap–water systems by Skoulios (31), who suggested that the reason for thisbehavior was that certain water layer thickness cannot exist in a gelphase. When the d spacings of the DGMS–distilled water gels are increasedup to 64 A, corresponding to a water concentration of 20%, the d valuesremain constant at 64 A, independent of the water concentration. The thick-ness of the bimolecular lipid layer da was calculated by analysis of theneutralized DGMS–water systems to be 55 A (22). The value of da of theDGMS–distilled water systems was found to be only approximately 50 A.The differences in thickness of the lipid bilayers may be due to a differentangle of tilt of the hydrocarbon chains toward the water layer in the twosystems. The da value of 55 A in the neutralized systems is identical to thelong spacing of the anhydrous a-form of DGMS. The water layer thickness,dw, is 175 A in mixtures containing 75% water. Undoubtedly, the swellingcontinues with a higher water content. The optical texture of neutralizedDGMS gels containing 95% water indicates a lamellar structure, and suchgels may have water layers of several hundred angstroms.

The microstructure of an a-gel phase and the corresponding b-crystal-line coagel (hydrate) is demonstrated in Fig. 20. The a-gel has a typicallamellar structure with lipid bilayers alternating with water layershaving a thickness of about 200 A. The coagel is a tridimensional networkof b-crystalline platelets suspended in water.

When a lamellar phase-of monoglyceridests cooled and a gel phase isformed, the x-ray spacings (d) increase, although the proportion between thewater content and the lipid phase is kept constant. This phenomenon is dueto the crystallization of the hydrocarbon chains in the lipid bilayer, whichresults in a decrease in the specific surface area (S) of the monoglyceridemolecules in contact with the water, as shown in Table 5. The thickness ofthe lipid layer da increases from 38 A in the lamellar phases to 55 A in the gelphases, due to the crystallization process, whereby the hydrocarbon chainsare stretched out and aligned parallel to each other. At the same time, thespecific surface area S decreases from 32 A2 in the lamellar phase to 22 A2 inthe gel phases. The crystallization of the hydrocarbon chains cannot aloneaccount for the increase in interplanar spacings that occur on transition


from the lamellar phase to the gel state. It can be seen from Table 5 that thewater layer thickness increases during the formation of a gel, although thetotal water content is kept constant. The main reason for this extension ofthe water layer is the decrease in specific surface, S, of the monoglycerides incontact with the water, which is, in turn, caused by a lateral contraction ofthe lipid molecules. The rearrangement of the molecules during the transi-tion from the lamellar phase to the gel phase is illustrated schematically inFig. 13, where the relative distances for d, da, and dw are shown for alamellar phase and a gel phase containing 40% water.

Sodium salts of stearic acid added directly to DGMS increase theswelling capacity of the gel phase to the same extent found when neutraliz-ing the free fatty acids with sodium hydroxide. The minimum concentrationof soaps needed to obtain the maximum degree of swelling is about 0.5% ofthe DGMS. The effect of sodium and calcium soaps on the gel phase is quitedifferent. Sodium stearate gives a high degree of swelling, whereas calciumstearate inhibits the swelling of a gel (23). The ionic active emulsifiers shownin Table 9 also increase the swelling and stability of monoglyceride gels,whereas the nonionic emulsifiers do not have this effect. Finally, it should be

Figure 20 Microstructure of (a) an a-gel phase of distilled, saturated mono-

glycerides in water (25 : 75) by freeze-fracture transmission electron microscopy

(bar¼ 0.2mm) and (b) the corresponding coagel phases (hydrate) by scanning

electron microscopy (bar¼ 2mm). The water layer thickness of the gel phase was

measured to approximately 200 A by x-ray diffraction analysis. The phase (b)

is a network of b-crystalline platelets of monoglycerides in water. (Courtesy of Dr. W.

Buchheim, Kiel, and Danisco, Denmark.)


Table 9 Influence of Various Additives on Swelling Capacity of DGMS and Distilled Water Systems Containing 70%

Distilled Water

Percent of

additives by

weight of DGMS

pH of

aqueous

mixtures

Lamellar phase (60�C) Gel phase (25�C)

Spacings

d (A)

Water

layer

thickness

dw (A)

Spacings

d (A)

Water layer

thicknessa

dw (A)

Nonionic additives

Tetraglycerol monostearate 4 4.7 54.1 16.0 62.3 12.3

Ethoxylated monoglycerides 4 4.6 54.6 16.5 79.1 29.1

Polyoxyethylene-(20)-sorbitan monostearate 4 4.2 61.7 23.6 67.0 17.0

Polyoxyethylene stearate 4 4.3 54.6 16.5 67.0 17.0

Lauryl alcohol polyglycol ether 1 4.3 55.1 17.0 61.7 11.7

Anionic active additives

Succinylated monoglycerides (sodium salt) 4 5.4 123.2 85.1 156.0 101.1

Sodium stearoyl-2-lactylate 4 4.7 128.4 90.3 163.0 128.1

Sodium lauryl sulfate 1 6.3 131.2 93.1 198.0 143.1

Sodium stearate 1 6.9 131.0 92.9 184.0 129.1

Calcium stearate 1 — 130.0 91.9 62.3 12.3

Cation active additives

Cetyl pyridinium bromide 1 4.6 131.0 92.9 192.0 137.1

Benzotonium chloride (Hyamin 1622) 1 4.6 128.0 89.9 60.5 10.5

Control

DGMSþ distilled water — 5.0 54.1 16.0 64.0 14.0

aCalculations based on da¼ 54.9 A for systems with ionic additives and da¼ 50 A for system with non-ionic additives.

Source: From Ref. 23.


mentioned that the stability of the gel phase is very sensitive to the ionconcentration in the water. The addition of 0.04% sodium chloride to thelamellar phase of monoglycerides in water inhibits the formation of a totallyswollen gel at a high water content.

The stability of the monoglyceride gel structure or other emulsifiers inaqueous systems is an important factor in preventing changes in the func-tional effect of emulsifiers when gels are used as aerating agents.

Lipophilic, a-tending emulsifiers, such as ACETEM, LACTEM, orPGMS, and glycerol mono-olein (GMO) can swell in their crystallinestate in the presence of water and form a-crystalline gel structures (32,33).This gel formation can be shown by x-ray diffraction analysis, as shown inTable 10. The emulsifiers, PGMS or GMO, are dissolved in coconut oil(Melting point¼ 31�C) and poured over a water phase in a beaker andthen cooled to 5�C for several hours. Samples of the interfacial layer andupper bulk phase free of water are taken for x-ray analysis.

The x-ray data on the water-free bulk phase show two sets of longspacings representing (A) the PGMS or GMO and (B) the coconut oil. Thisshows that the emulsifiers do not cocrystallize with the coconut oil but formcrystalline aggregates separately from the coconut oil crystals.

The interfacial layer also exhibits two sets of long spacings. The valuesof 56.1 A and 66.0 A represent long spacings of PGMS water gel and GMOwater gel respectively, and the values of 35.4–36.1 A represent long spacingsof the coconut oil.

An interesting feature is the increase in the long spacings of the emul-sifier phase (A) from the bulk system to the interfacial layer. For PGMS, theincrease is about 7 A, and for GMO, the increase is 17 A. The changes inlong spacings induced by the contact with water can only be due to water

Table 10 X-ray Diffraction of Interfacial Layers from Coconut

Oil–Emulsifier–Water System at 5�C

Sample preparation

Long spacings d (A)

Short spacings

d (A)

A

Emulsifier

B

Coconut oil

Interfacial layer

Coconut oil–PGMS–water 56.1 35.4 4.18–3.83

Coconut oil–GMO–water 66.0 36.1

Bulk phase

Coconut oil–emulsifier (1 : 1)

(melt, cooled to 5�C) 48.9 36.3 4.29–4.16–3.97–3.82–3.65


penetration through the polar regions of the emulsifier crystals driven by ahydration force, forming an a-gel structure. The hydration of GMO atambient temperature or even lower temperatures is in good agreementwith its phase diagram.

The a-gel formation is of great importance in whippable emulsions(creams, toppings), and the hydration of unsaturated monoglycerides at lowtemperature can be used to make powdered monoglyceride blends that aredispersible in water and used as cake emulsifiers.

F. Liquid Crystals and Emulsion Stability

Liquid crystals in emulsions were observed in 1969 by Friberg and hisco-workers (34), and in 1971 (35), they demonstrated that liquid-crystallinephases may form on the surface of oil droplets in o/w emulsions and providestability against coalescence.

The presence of liquid-crystalline mesophases in food systems is nota common feature partly due to a low concentration of polar emulsifiersand partly due to lack of stability of emulsifier–water mesophases, whichmay be negatively affected by other food ingredients, such as fats, proteins,carbohydrates, salts, and so forth.

An example of o/w emulsions where liquid crystals have been found bypolarized microscopy is salad dressings made with egg yolk lecithins. Amicrograph of such an emulsion is shown in Fig. 21. The strong birefringentlayers around the oil droplets are present due to a formation of liquid-crystalline layers. Similar observations have been made with other typesof emulsions (36,39).

The existence of lamellar, interfacial multilayers has been shown toprovide stability to emulsions of sunflower oil in water by Pilpel andRabbani (37,38). The emulsifiers used were sorbitan monopalmitate(Span-40) in combination with Polysorbate 40. The optimal ratio of thetwo emulsifiers for emulsion stability was found to be 5mol of Span-40 to1mol of Polysorbate 40. With this ratio of the lipophilic Span-40 and thehydrophilic Polysorbate 40, ideal conditions for the formation of liquid-crystalline lamellar structure at the o/w interface exist. The formationof such lamellar structures around oil droplets was shown by electronmicroscopy.

G. Polar Lipid–Water Phases of Natural Origin

Plants contain polar and nonpolar lipids, and the polar lipids may exist inthe form of liquid-crystalline phases containing water. This is the case incereals (e.g., wheat flour), which contains about 2.5% total lipid on dry


basis. The polar lipids in wheat flour are glycolipids, phospholipids, freefatty acids, and monoglycerides, which amount to approximately 45% ofthe total wheat lipids.

Carlson et al. (40) demonstrated that both lamellar and hexagonal IIliquid-crystalline phases exist in ternary systems of nonpolar and polarwheat flour lipids and water. The amount of lamellar lipid water phaseswas found essential for the bread-baking quality of the flour. Liquid-crystalline phases of polar lipids and water have also been found inmature wheat endosperm by freeze-fracture electron microscopy (41).

IV. FUNCTIONAL PROPERTIES OF EMULSIFIERS IN FOODS

A. Emulsification and Emulsion Stability

The manufacture of o/w food emulsions often involves a homogenizationprocess under turbulent flow conditions. The relative contribution to dro-plet disruption by energy input (e.g., homogenisation pressure) and theeffect of emulsifiers on reduction of interfacial tension between oil and

Figure 21 Commercial salad dressing made with egg yolk lecithin. The oil droplets

are seen as dark spheres surrounded by liquid-crystalline layers of phospholipids,

showing a strong birefringence in polarized light microscopy. Bar¼ 100mm.


water is approximately 100 : 1. The emulsification process and final particledistribution of the emulsion are thus mainly controlled by the energy input,and the influence of emulsifiers is negligible.

In the case w/o emulsions (e.g., margarine, spreads), which is madeunder low-energy impeller emulsification methods, the addition of emulsi-fiers has a significant effect on reducing the water droplet size.

B. Interfacial Emulsifier–Protein Interactions in Emulsions

Many food emulsions contain milk proteins or proteins from plants oranimal origin. The main function of emulsifiers in such emulsions is eitherto increase stability toward coalescence or creaming/precipitation duringlong-term storage (e.g., recombined milk, coffee whiteners, salad dressings,etc.) or to induce destabilization and increase whippability of emulsions tobe aerated (e.g. ice cream mix, cream, toppings, etc.).

Interfacial interactions between emulsifiers and proteins in emulsionsplay a vital role for the physical properties of emulsions. A cooperativeemulsifier–protein adsorption at the surface of fat globules may improvethe surface film strength and coherence, increasing emulsion stability (17).This is found with anionic emulsifiers (DATEM, SSL) that bind proteins tothe surface of fat or oil droplets via a complex formation (42). Anionicemulsifiers are therefore very effective in emulsions, where long-term stabi-lity is important.

In contrast, nonionic emulsifiers (polysorbates, monoglycerides,LACTEM, or PGMS) compete with proteins for interfacial adsorptionand may displace the proteins more or less from the interface. The displace-ment of protein from fat globule surfaces is part of the destabilization ofwhippable emulsions. Other important functions of emulsifiers in the desta-bilization process is to enhance crystallization of the fat phase and to reducethe coherence and viscoelastic properties of the surface film, making the fatglobules more sensitive to shear (43,44), increasing flocculation and partialcoalescence. The formation of fat globule aggregates during whipping isenhancing the foam formation and stabilizes the foam structure of aeratedemulsions (17,33,45,46).

C. Interaction with Starch Components

Food emulsifiers are used in many starch-based foods such as bakery pro-ducts, processed potatoes, breakfast cereals, pasta foods, and so forth. Theprimary function of emulsifiers such as monoglycerides, SSLs CSLs, andothers is to form a water-insoluble complex with the starch component


amylose. The amylose is present in most starches in an amount of 17–25%and is the source of textural problems (stickiness, firmness) in starch-con-taining foods. By formation of a lipid–amylose complex, the texture ofreconstituted potato products and pasta foods is improved.Monoglycerides in wheat bread are used to reduce crumb firmness anddelay the staling process. Saturated monoglycerides are the most effectiveamylose-complexing agents and are used in bread in the form of the b-crys-talline hydrates (47), or dispersible fine powders. Stearoyl lactylates (SSLs,CSLs) are also used as starch-complex agents in many products, primarilybread.

The amylose-complex formation in wheat bread may also affect theretrogradation of amylopectin, the major starch component. At low use1evels of monoglycerides (e.g., 0.3–0.5% based on flour weight), the mono-glyceride mainly interacts with the amylose fraction. At higher concentra-tions of monoglyceride (1–2%), they seem to interact directly with theamylopectin fraction, reducing the rate of retrogradation, which is partlyresponsible for the bread-firming process during storage (47).

D. Dough-Strengthening Effects of Emulsifiers

Hydrophilic emulsifiers such as DATEMs, SMGs, SSLs, CSLs, and poly-sorbates are used in yeast-raised doughs to increase shock-resistancefermentation stability and enhance volume of baked products such asspeciality bread, fiber bread, rolls, and buns. The emulsifier functionmechanism in doughs is not completely understood, although many emulsi-fier action models have been proposed. A relation to the physical propertiesof natural flour lipids may be relevant because it is known that the polarflour lipids are very important for the baking quality of flour (48). The polarflour lipids form liquid-crystalline phases in water and—especially thelamellar phase formed by polar flour lipids, such as galactolipids andphospholipids—are considered important for the baking quality of flour.Polar-lipid bilayer structures may associate with hydrated gliadin proteins atgas cell–water interfaces and improve the gas-retention capacity of thedough during fermentation.

All effective dough-strengthening emulsifiers (DATEMs SSLs,polysorbates, and lecithins) form lamellar liquid-crystalline phases inwater under conditions similar to dough mixing. They may thus combinewith the polar lipids in flour into biomembrane like structures andsupport the function of the native, polar lipids in flour. The role oflipids in bread making is extensively discussed by Eliasson andLarsson (49).


E. Legal Aspects

Before a food emulsifier is permitted for use in foods, it has to betested in a number of toxicological studies including short- and long-term feeding trials on several animal species together with studies onmetabolism.

The evaluation of emulsifiers is done by the Joint Expert Committeeon Food Additives (JECFA) of the Food and Agriculture OrganizationWorld Health Organization (FAO/WHO), by the commission of theEuropean Communities’ Scientific Committee for Food (SCF), and in theUnited States by a department of the Food and Drug Administration(FDA).

The results of the scientific, independent toxicological evaluations areused by national health authorities for food regulations.

The acceptable daily intake (ADI) values and regulatory identificationnumbers for the emulsifiers dealt with here are shown in Table 1. All emul-sifiers approved by local health authorities should be considered as safe foodingredients when used within their limits of the ADI values, which aregiven in the reports of the Joint FAO/WHO Expert Committee on FoodAdditives (50).

V. CONCLUDING REMARKS

The aim of this chapter is to present a general review of the most commonlyused emulsifiers in the food industry. Food emulsifiers should be consideredpolar lipids with specific properties and functions. Their chemical composi-tion determines their physical properties, which is of importance for theirdifferent functions in foods. Most interactions between emulsifiers and otherfood ingredients take place in surface monolayers or at mixed lipid–proteininterfacial structures. Such interactions takes place in emulsions, resultingin increased stability of some types of emulsions and partial destabilizationof others.

The capability of lipid emulsifiers to form association structures withwater in form of lamellar dispersions or vesicles enhance molecular interac-tions of emulsifiers with water soluble food components (e.g., starch com-ponents). Furthermore, specific lipophilic emulsifiers may be incorporatedin the crystal lattice of fats, thus controlling crystallization and texture of theproduct.

The molecular behavior of emulsifiers in all states or order, in bulkphases and at interfaces in emulsions, foams, and colloidal systems is veryimportant to understand in order to be able to predict functional propertiesof emulsifiers in foods.


REFERENCES

1. R. O. Feuge and A. E. Bailey, Oil Soap 23, 259 (1946).

2. N. O. V. Sonntag, in Bailey’s Industrial Oil and Fat Products (P. Swern, ed.),

Wiley, New York, 1982, Vol. 2, p. 99.

3. J. P. Brandner and R. L. Brikmeier, J. Am. Oil Chem. Soc. 37, 390 (1960).

4. J. B. Søe, Fette, Seifen, Anstrichtmittel 85, 72 (1983).

5. K. Larsson, in Lipids—Molecular Organization, Physical Functions and

Technical Applications, Oily Press, Dundee, Scotland, 1994, p. 27.

6. J. B. Martin and E. S. Lutton, J, Am. Oil Chem. Soc. 49, 683 (1972).

7. R. O. Feuge, E. J. Vicknoir, and N. V. Loogren, J. Am. Oil Chem. Soc. 29,

11 (1952).

8. G. Schuster and W. Adams, Seifen, Ole, Fette, Wachse 3, 61 (1981).

9. J. Holstborg, B. V. Pedersen, N. Krog, and S. K. Olesen, Colloids Surfaces B:

Biointerf. 12, 383 (1999).

10. E. S. Lutton, C. B. Stewart, and J. B. Martin, J. Am. Oil Chem. Soc. 49, 186

(1972).

11. F. R. Benson, in Nonionic Surfactants (M. J. Schick, ed.), Marcel Dekker,

New York, 1987, p. 247.

12. L.I. Osipow and W. Rosenblatt, J. Am. Oil Chem. Soc. 44, 307 (1967).

13. B. F. Szuhaj and G. R. List, eds., Lecithins, American Oil Chemical Society,

Champaign, IL, 1985.

14. D. M. Small, in The Physical Chemistry of Lipids (D. M. Small, ed.), Plenum,

New York, 1986, p. 386.

15. N. Krog, in Crystallisation Processes in Fats and Lipid Systems (N. Garli and

K. Sato, eds.), Marcel Dekker, New York, 2001, p. 505.

16. D. M. Small, in The Physical Chemistry of Lipids (D. M. Small, ed.), Plenum,

New York, 1986, p. 475.

17. N. Krog, in Ingredient Interactions—Effects in Food Quality, 2nd ed. (A. G.

Gaonkar, ed.), Marcel Dekker, New York, in press.

18. O. Lehman, Flussige Kristalle, Engelmann, Leipzig, 1904.

19. V. Luzzati, H. Mustacchi, A. Skoulios, and F. Husson, Acta Crystallogr. 13,

660 (1960).

20. G. J. T. Tiddy, Phys. Rep. 57, 1 (1980).

21. H. Hauser, in Proc. Eur. Sci. Found. Workshop, 4th (P. L. Luisi and B. E.

Straub, eds.), 1984, p. 37.

22. K. Larsson, Zeit, Phys. Chem. Neue Folge 56, 173 (1967).

23. N. Krog and A. P. Borup, J. Sci. Food Agric. 24, 691 (1973).

24. S. T. Hyde, S. Andersson, B. Ericsson, and K. Larsson, Z. Kristallogr. 168, 213

(1984).

25. N. Krog and K. Larsson, Chem. Phys. Lipids 2, 129 (1968).

26. B. Ericsson, Interactions between globular protein and polar lipids, Thesis,

Lund University, Sweden, 1986, p. 20.

27. K. Larsson, British patent No. 1,174,672, 1967.

28. W. Hemker, J. Am. Oil Chem. Soc. 58, 114 (1981).


29. F. Reiss Husson, J. Mol. Biol. 25, 363 (1967).

30. K. Larsson and N. Krog, Chem. Phys. Lipids 10, 177 (1973).

31. A. Skoulios, Adv. Colloid Inferf. Sci. 1, 79 (1967).

32. J. M. M. Westerbeek and A. Prins, in Food Polymers, Gels and Colloids

(E. Dickinson ed.), Royal Society of Chemistry, Cambridge, 1991, p. 147.

33. N. Krog, in Emulsions—A Fundamental and Practical Approach (J. Sjoblom

ed.), Kluwer, Dordrecht, 1992, p. 61.

34. S. Friberg, L. Mandell, and M. Larsson, J. Colloid Interf. Sci. 29, 155 (1969).

35. S. Friberg and L. Rydhag, Kolloid-Z. Z. Polym. 244, 233 (1971).

36. K. Shinoda and S. Friberg, Emulsions and Solubilization, Wiley, New York,

1986, p. 159.

37. N. Pilpel and M. E. Rabbani, J. Colloid Interf. Sci. 119, 550 (1987).

38. N. Pilpel and M. E. Rabbani, J. Colloid Interf. Sci. 122, 266 (1988).

39. S. Hamden, K. Anuar, and G. R Michell, J. Am. Oil Chem. Soc. 72, 109 (1995).

40. T. L.-G. Carlson, K. Larsson, Y. Miezis, and S. Poovarodom, Cereal Chem. 56,

417 (1979).

41. A. Al Saleh, D. Marion, and D. J. Gallant, Food Microstruct. 5, 131 (1986).

42. E. Dickinson and Soon-Taek Hong, J. Agric. Food Chem. 42, 1602 (1994).

43. E. Davies, E. Dickinson, and R. D. Bee, Ind. Dairy J. 11, 827 (2001).

44. E. Dickinson, R. K. Owusu, and A. Williams, J. Chem. Soc. Faraday Trans. 89,

865 (1993).

45. N. Krog, N. M. Barfod, and W. Buchheim, in Food Emulsions and Foams

(E. Dickinson, ed.), Royal Society of Chemistry, London, 1987, p. 144.

46. H. D. Goff, in Lipid Technologies and Applications (F. D. Gunstone and F. B.

Padley, eds.), Marcel Dekker, New York, 1997, p. 329.

47. N. Krog, S. K. Olesen, H. Toernaes, and T. Joensson, Cereal Food World 34,

281 (1989).

48. F. MacRithchie, J. Sci. Food Agric. 28, 53 (1977).

49. A.-C. Eliasson and K. Larsson, Cereals in Bread Making, Marcel Dekker,

New York, 1993.

50. World Health Organization, Technical Report Series Nos. 539 Toxicological

evaluation of certain food additives with a review of general principles and of

specifications (1974), 653 Evaluation of certain food additives (1980), and 683

Evaluation of certain food additives and contaminants (1982), Geneva, 1978.


3Molecular Organization in Lipidsand Emulsions

Kare LarssonCamurus Lipid Research, Lund, Sweden

I. INTRODUCTION

Important knowledge on the lipid structure in emulsions is based on studiesof simple systems containing only a few components, such as a ternarysystem consisting of a polar lipid, a triglyceride oil, and water. Structuresand phase equilibria in such model system provide information on emulsion-stabilizing mechanisms which can be translated into the behavior of complexfood emulsions. The structures in the different states of order of lipid andlipid–water phases will be summarized with special regard to emulsions.Additional information on the chemistry and physics of emulsions arefound in Ref. (1).

Lipids are unique in their ability to form a wide variety of structures,ranging from micellar solutions to liquid-crystalline phases and differentcrystalline forms (polymorphism). An important factor behind the molecu-lar organizations they exhibit is the amphiphilic character of the molecule—the possibility to orient hydrophilic groups toward polar regions, such aswater, and hydrophobic groups separated into nonpolar regions.

II. THE SOLID STATE

In crystals, it is possible to obtain exact and detailed information on themolecular conformation as well as the lateral molecular packing.


Knowledge about lipid crystal structures is therefore of fundamental impor-tance for discussions of the structure in disordered states, such as liquid-crystalline lipid–water phases. Furthermore, there is a close resemblancebetween crystal structures and the structures of lipid monolayers at interfaces(e.g., on a water surface). A remarkable property of lipid crystalline phasesthat should be mentioned in this connection is that complex mixtures likenatural fats can form solid solution with wide variations in molecular size.

A. Experimental Methods

When single crystals are available, it is possible to perform a completestructure determination by the x-ray diffraction technique. Even if singlecrystals are not available, a considerable amount of information on solid-state packing can be obtained by powder diffraction. It is thus possible toobtain a so-called long spacing from which the main features of the arrange-ment of the molecules in layers can be obtained, such as the number ofmolecular layers in the repetitive unit layer and the angle of tilt of themolecules within each layer. In the wide-angle region of the diffractionpattern, it is possible to identify dominating so-called short-spacing lines,which, in many cases, are sufficient for identification of the hydrocarbonchain packing.

X-ray diffraction is by far the most powerful method used to determinethe structures in the solid state, even if important complementary informa-tion has been obtained by other methods (e.g., by infrared spectroscopy) (1).

B. Molecular Orientation in Lipid Crystals

The characteristic feature of lipid crystal structures is the arrangement ofmolecules in layers with the polar groups at the surface of the layer, asshown in Fig. 1. The molecular layers usually form unit layers of doublemolecular length, lipid bilayers, in order to permit polar interaction betweenthe molecular layers (Fig. 1a). This means that there are only relatively weakvan der Waals attractive forces between these bilayers separated by themethyl end-group gap. This explains the crystal morphology; the crystalsform very thin plates parallel to these planes, with a pronounced tendencyfor cleavage at the methyl gap interface.

In cases with very weak attractive forces between the polar headgroups or when the polar groups are not end groups, a head-to-tail arrange-ment as shown in Fig. 1b can be adopted. An example is ethyl stearate (2), inwhich the polar group is too far from the end of the linear molecule to givedimerization. Methyl stearate (3), on the other hand, forms the usualhead-to-head structure.


A head-to-tail arrangement within each layer, as shown in Fig. 1c, hasbeen observed in cases with very bulky polar head groups. By mean of sucha structure, the polar groups obtain a cross-sectional area twice as large asthe hydrocarbon chains.

In cases where the molecules are tilted, they are usually all parallel. Thetilt results in an increased space for the polar head group (see Section II. F).In some structures, however, the angle of tilt alternates between oppositedirections in successive layer, as indicated in Fig. 1d. Such an arrangementpossesses particular advantages in cases of solid-state phase transitions, astranslations along the methyl end-group planes can be avoided. Finally, itshould be mentioned that structures with two chain directions within eachmolecular layer have been observed in cases where the forces between thepolar head groups are of dominating importance (e.g., in amides) (4).

C. Hydrocarbon Chain Packing

The hydrocarbon chains are always in the extended planar zigzag conforma-tion in lipid crystal forms. There are a few alternative close-packing arrange-ments of such chains, which can be best described by the corresponding

Figure 1 Schematic of the major types of molecular orientation in layers: (a) head-

to-head arrangement, all molecules directed in the same way within each layer;

(b) head-to-tail arrangement, all molecules directed in the same way within each

layer; (c) head-to-tail arrangement within each layer; (d) tilted molecules within mono-

molecular layers (alternating directions of the tilt, as shown here, or nonalternating).


subcell (the smallest repetition unit within the layer). All chain packingalternatives that have been observed are summarized in Fig. 2. The triclinicpacking, T parallel, and the orthorhombic one, O perpendicular, representthe most efficient packing from an energetic point of view, and they are

Figure 2 Hydrocarbon chain packing alternatives as defined by the corresponding

subcell. One zigzag period is seen in the direction of the hydrocarbon chains with the

atomic positions in the chain direction given by their fractional coordinates. Open

circles represent hydrogen atoms and filled circles represent carbon atoms.


observed in n-paraffins, for example. All chain planes are parallel in thetriclinic packing, whereas every second chain plane is perpendicular to therest in the orthorhombic packing. The monoclinic packing, M parallel, hasbeen observed in racemic 1-monoglycerides (5). An orthorhombic subcell,O0 perpendicular, closely related to the common one, O perpendicular, wasfound in a branched fatty acid (6). In other structures with disturbances inthe hydrocarbon chain region, two more orthorhombic chain packing typeshave been observed: O parallel and O0 parallel, both with all chain planesparallel (7,8).

An interesting disordered crystal form, usually termed the alpha form(9), is observed near the melting point of many lipids. Its x-ray single-crystaldata are in agreement with a hexagonal lateral symmetry of the hydrocar-bon chains with rotational freedom (10). Due to the rotation of the mole-cules and its physical properties, this form is in fact a plastic crystal. In somelipids, however, the chains are anchored at the polar groups (e.g., in trigly-ceride), and it is obvious that only oscillation movements of the chains arepossible. Finally, it should be mentioned that this hydrocarbon chainarrangement occurs in lipid–water phases and even in emulsions.

D. Polymorphism

The multiple melting phenomena exhibited by triglycerides was firstexplained by Malkin (11) as due to the occurrence of an alternative crystalform: polymorphism. One possibility for polymorphism is the arrangementof the different hydrocarbon chain close-packing types described earlier.The most common reason for polymorphism, however, is the possibilityfor variations in the angle of tilt of the hydrocarbon chains. The chainsare successively displaced one or more whole zigzag periods in relation toadjacent chains so that the angle of tilt toward the end group planeis increased or decreased and the hydrocarbon chain region will not bechanged.

Most lipids exhibit polymorphism. This can be illustrated by fattyacids. An even n-fatty acid in the chain-length range C12 to C18 can beobtained in three different crystal forms: the A, B and C forms. The crystalforms B and C have the same packing (O perpendicular) with different anglesof tilt, whereas the A form shows another chain packing (T parallel).

E. Fatty Acids

The structure shown in Fig. 1a is characteristic for fatty acids with normaland saturated hydrocarbon chains. Branches can be accommodated into the


hydrocarbon chain region by relative displacements of adjacent chains or bychanges in the tilt direction at the branch position (12).

When there are double bonds in the hydrocarbon chain, they willaffect the chain packing differently according to the bond configuration.In oleic acid, where there is a cis double bond, there is a change in the chaindirection at the double bond, whereas the corresponding trans isomer,elaidic acid, shows only one chain direction (8). The crystal structure ofoleic acid is shown in Fig. 3.

Figure 3 Crystal structure of oleic acid viewed along the b axis. (From Ref. 8.)


F. Monoglycerides and Emulsions Based on CrystallineMonoglycerides

Monoglycerides are most commonly used emulsifier in foods (see Chapter2). An emulsification process, involving cooling through the crystallizationof the monoglyceride solved in the oil, will result in the formation of acrystalline phase at the oil–water interface, as shown both from surfacetension studies and phase identification (13). Organized lipid phases at theoil–water interface is a common emulsification-stabilizing mechanism infood emulsions; we will come back to this in connection with liquid crystals.

The crystal structure of a saturated 1-monoglyceride is shown in Fig. 4.There is a complex hydrogen-bond system linking the molecules together intwo dimensions. When saturated monoglycerides are used as emulsifiers, afinal emulsion product can be obtained stabilized by a crystalline monogly-ceride film at the interface, with a methyl end group surface toward the oilphase and a monoglycerol head group surface toward water.

G. Diglycerides and Triglycerides

Crystal structure work has shown that there are two types of molecular con-formation in diglyceride crystals. In one alternative, the two hydrocarbon

Figure 4 Crystal structure of a 1-monoglyceride viewed along two perpendicular

directions. Also, 2-monoglyceride form similar to complex hydrogen bond networks

linking the molecules in two dimensions. (From Ref. 5.)


chains have the same direction in relation to the head group, and thenbilayers are formed just like the situation in monoglycerides. The otheralternative is somewhat similar to triglyceride structures, with both chainspointing in opposite directions. Such a monolayer is then a unit layer in thecrystal.

Simple saturated triglycerides crystallize in layers where extendedmolecules are arranged in the pairs as shown in Fig. 5. The acyl chainsin the 1- and 3-position are extended, forming one chain, whereas thechain in the 2-position forms a branch, This is the characteristic molecularconformation of all triglycerides in the solid state. The presence of cisdouble bonds will result in a chain-sorting tendency so that unsaturatedchains are localized in the same layer. This, in turn, will mean thatsometimes the branching chain can be formed by the chain in the 1- or3-position. Furthermore, the double-chain unit layer as shown in Fig. 5cannot allow ideal chain sorting, and a triple-chain layer is instead theunit layer.

Simple triglycerides (all chains equal) show three polymorphicforms, and even complex triglycerides have a similar behavior althoughmore forms may occur. The nomenclature now generally accepted toidentify the polymorphic forms is based in x-ray diffraction with thefollowing criteria:

Alpha form: One dominating diffraction line at 4.2 A.Beta prime form: Two diffraction lines at 4.2 and 3.8 A dominate.Beta form: Name for all other forms.

The crystal structure of the beta prime from and mechanisms involvedin the polymorphic transitions have recently been determined by thepioneering work by Sato and co-workers (cf. Ref. 14).

The different polymorphic forms have different morphologies and thismay influence emulsion stability. The beta prime form, for example, formsthin needles and they can sometimes extend through the oil–water interfaceand result in aggregation phenomena.

III. LIQUID-CRYSTALLINE PHASES

This term was introduced in order to describe structures with partialdisorders like in liquids. Lipid–water phases form a wide variety ofliquid-crystalline structures with technical significance, such as in foodemulsions.

A requirement for the formation of liquid-crystalline phases is thatthe hydrocarbon chains are disordered, like in the liquid state of paraffin.


Figure 5 Molecular arrangement of trilaurin in the beta crystal form viewed along

the shortest unit-cell axis. (From Ref. 5.)


The fundamental work in this field was reported more than four decadesago by Luzzati and co-workers (cf. Ref. 15). The different structures will bedescribed and complementary information is found in Chapter 4.

A. Methods

The most informative method is x-ray diffraction (or x-ray scattering). Dueto the long unit periods in these phases, it is necessary to use small-anglediffraction/scattering equipment. The identification of liquid-crystallinephase is straightforward and based on the existence of sharp diffractionlines corresponding to long spacings in the structure combined with nodiffraction lines in the small-angle region, only a diffuse line at 4.5 A.This line is due to the disordered chains, and the same line is obtainedfrom a liquid triglyceride oil.

Spectroscopic methods, nuclear magnetic resonance (NMR) in partic-ular, can also be used for phase identification and, in addition, provideimportant complementary information on dynamic properties.

Phase diagrams are useful tools in order to describe lipid–watersystem, and examples relevant to foods are shown in Chapter 9. Sampleswith different water contents are equilibrated and single phases are analyzedby x-ray diffraction. First, inspections should be performed in the polarizingmicroscope, and the birefringence is an important feature of the liquid-crystalline phases. Only the cubic phases lack birefringence, but they areeasy to detect because of their high viscosity.

B. Lamellar Liquid Crystals

This phase is characterized by a series of diffraction spacings in the ratio1 : 2 : 3 : 4. . . corresponding to a one-dimensional repetition with the first-order spacing d equivalent to the thickness of the water layer and thebimolecular lipid layer. By simple geometry, the lipid bilayer thicknessand the surface area per polar group can be calculated. The structure isshown in Fig. 6.

A special structure related to the lamellar liquid-crystalline phase isthe so-called gel phase. It also consists of water layers alternating withlipid bilayers, but the hydrocarbon chains are crystalline. This phase isoften obtained as a meta-stable state when the lamellar liquid-crystallinephase is cooled below the transition point to crystalline chains. Thechains in this phase are usually arranged according to the hexagonalsymmetry, indicating rotational or oscillation movements around thechains axes.


C. Hexagonal Liquid Crystals

A common aqueous lipid phase is often formed in lipids with a small hydro-carbon chain region at about equal amounts of water and lipid. It wascalled the middle phase in the earlier literature on soaps. This phaseshows diffraction patterns with the square roots of the spacings in theratios 1 : 3 : 4 : 7 : 12. . ., the condition of a two-dimensional hexagonal lattice.The swelling in water corresponds to an ideal two-dimensional swelling. Thestructure is shown in Fig. 6. Very polar lipids, such as lyso-phospholipids,form infinite cylinders with a liquid hydrocarbon chain core arranged in ahexagonal array and the polar heads facing the outside water. The spacebetween the cylinders is taken up by water.

An alternative hexagonal structure of reversed type (i.e., water cylin-ders surrounded by lipid molecules) has been observed in many lipid systemswith a relatively large hydrocarbon chain proportion. This phase, alsoshown in Fig. 6, is identified in the same way from x-ray diffraction pattern.By simple geometry, the diameter of the cylinders can be calculated fromthat the molecular length and the surface area per polar group. There isusually no problem in determining which is the true alternative of thesetwo hexagonal structures, once the molecular dimensions have beencalculated.

D. Cubic Liquid-Crystalline Phases

Cubic phases are the most complex of lipid–water phases and their bicon-tinuous structures have recently been revealed. Such a phase must showdiffraction patterns where the square roots of the spacings form the ratios1 : 2 : 3 : 4 : 5 : 6 : 8 :9 : 10 : 11 : 12. . . . Absent lines are related to the actualsymmetry (space group). The structure of one type of a cubic lipid–water

Figure 6 A fragment of the lamellar liquid-crystalline phase (in the middle), the

hexagonal (to the right) and the reverse hexagonal (to the left) liquid-crystalline

phases.


phase is shown in Fig. 7a. It has been shown that the lipid bilayer in such aphase follows an infinite, periodic, minimal surface (16). A minimal surfacehas zero average curvature in all points; that is, it is as convex as it isconcave. The structure unit is illustrated in Fig. 7a, as well as the repetitionof these units into an infinite three-dimensional surface. The lipid bilayer iscentered on the surface, and on both sides, there is a water channel system.

Figure 7 Structure of the simplest bicontinuous cubic lipid–water phase (space

group Im3m) illustrated in (a) from the structure unit (plastic model) and the linkages

of these bilayer units. Another phase (space group Ia3d ) is shown in (b).


Six water channels meet in the structure unit. The lipid bilayer is free fromintersections and there is no connection between the two sides of the bilayer.

There are three different types of fundamental cubic minimal surface.They have different space groups, and x-ray data corresponding to allof them have been observed (16). All three have, in fact, been seen in thefood-grade emulsifier monoolein. A second type of structure is shown inFig. 7b, with three water channels joined into a structure unit. Finally, thereis a structure with four channels joined in tetrahedral geometry.

E. Liquid-Crystalline Disperson of the Lamellar Phaseand Emulsions

The lamellar liquid-crystalline neat phase can swell only up to a particularwater-layer thickness, which depends on the nature of the lipid. The situa-tion that will be considered here is the one occurring in most lipids, when adisperson of particles in an aqueous environment is formed when water isadded above the swelling limit of the lamellar phase. The lamellar phase inneutral lipids swells up to a water-layer thickness of about 20 A. The par-ticles always have a closed structure with concentric bimolecular layersalternating with water layers. Such particles (termed liposomes) can encap-sulate oil.

Friberg and co-workers more than three decades ago were the first torealize the significance of the lamellar liquid-crystalline phase in emulsionstabilization as demonstrated in the first edition of Food Emulsions. Thisstability criterium can easily be defined from phase equilibria; the oil phaseand the water phase are in equilibrium with a lamellar liquid-crystallinephase.

A phase transition mechanism might also be utilized for destabiliza-tion purposes. If heating of the lamellar phase, for example, results in theformation of a reverse phase, this transition at the surface of the emulsiondroplets will result in flocculation and coalescence.

Colloidal dispersions of cubic phases, cubosomes, can be preparedfrom food-grade components and may have applications in foods (cf.Ref. 17). In this connection, it should be mentioned that food additivesthat induce formation of a cubic phase with water has applications in orderto obtain aggregation of fat globules (e.g., in artificial cream products).

REFERENCES

1. D. Gunstone, J. L. Harwood, and F. B. Padley eds., The Lipid Handbook, 2nd

ed., Chapman & Hall, London, 1992.


2. S. Aleby, Acta Chem Scand. 17, 221 (1968).

3. S. Aleby and E. von Sydow, Acta Crystallogr. 13, 487 (1960).

4. J. D. Turner and E. C., Lingafelter, Acta Crystallogr. 16, 404 (1963).

5. K. Larsson, Arkiv Kemi 23, 29 (1964).

6. S. Abrahamsson, Acta Crystallogr. 12, 304 (1959).

7. E. von Sydow, Acta Chem. Scand. 12, 777 (1958).

8. S. Abrahamsson and I. Ryderstedt-Nahringbauer, Acta Crystallogr. 15, 1261

(1962).

9. K. Larsson, Acta Chem. Scand. 20, 2255 (1966).

10. K. Larsson, Nature 213, 383 (1967).

11. T. Malkin, Progress in the Chemistry of Fats and Other Lipids, Pergamon Press,

London, 1954, Vol. 2, p. 1.

12. S. Abrahamsson, Arkiv Kemi 14, 65 (1959).

13. N. Krog and K. Larsson, Fat Sci. Technol. 94, 55 (1992).

14. K. Sato, M. Goto, J. Yano, K. Honda, D. R. Kodali, and D. M. Small, J. Lipid

Res. 42, 338 (2001).

15. V. Luzzati, H. Mustacchi, A. Skoulious, and F. Husson, Acta Crystallogr. 13,

660 (1960).

16. K. Larsson, J. Phys. Chem. 93, 7304 (1989).

17. K. Larsson, Curr. Opin. Colloid. Interf. Sci. 5, 64 (2000).


4Interactions Between Proteins andPolar Lipids

Tommy NylanderLund University, Lund, Sweden

I. INTRODUCTION

Proteins and polar lipids coexist in biological systems, sometimes unasso-ciated with each other, but also as composite structures with specific actions(e.g., cell membranes and blood lipoproteins). They have a very importantphysical property in common—an inherent amphiphilic nature, which pro-vides the driving force for formation of associated structures of lipids as wellas for the folding of a polypeptide chain to form the unique conformation ofa native protein. An important consequence of this dualistic character is thatthe molecules will orient at an interface. Composite biological structures areformed in this way, such as cell membranes, which separate one microenvi-ronment from another. Therefore, it is not surprising that the major concernof the vast number of publications dealing with protein–lipid interactionshave been focused on biological membranes. The main concern has been togain a better understanding of biomembrane function, fusion, and othermembrane processes, which has evoked a number of studies of these com-plex lipid–protein ‘‘supramolecular’’ systems. There are a number of reviewson this expanding area of research, such as Refs. 1–5 and the book edited byWatts (6). The classic technique for studying protein–lipid interactions inthis context has been to study the properties of lipid or mixed lipid–proteinmonolayers at the air–aqueous interface by the surface film balance by itselfor in combination with techniques such as surface rheology measurementsand imaging techniques (4,7–14).


Many proteins have also the biological role of transporting moleculeswith hydrophobic properties, which bind to hydrophobic pockets in theprotein. One such protein is the major whey protein in milk, b-lactoglobu-lin—a lipocalin type of protein. The biological role of b-lactoglobulin isthought to be to transport retinol (or possibly other hydrophobic ligands)(15–19). Several studies have clearly demonstrated that this protein has ahigh-affinity binding site for a range of ligands like phospholipids, fattyacids, cholesterol, and triglycerides (19–26). This has implications for theprocessing of milk and dairy products in terms of, for instance, thermalstability, as discussed by Sawyer et al. (27).

Water-soluble polar lipids and synthetic analogs (surfactants) arewidely used in many technical and analytical applications. As we willdiscuss in the following, the interaction between proteins and polarlipids can lead both to stabilization of the protein structure and tounfolding. An example of the latter is the unfolding induced by sodiumdodecylsulfate (SDS), used for analytical purpose in the SDS–poly-acrylamide gel electrophoretic (SDS–PAGE) determination of molecularweights (28,29).

Proteins and lipids, separately as well as in mutual interaction, con-tribute significantly to the physical properties of many systems of techno-logical interest (e.g., emulsions and foams). Further, protein structure, andthereby function and properties, are affected in various ways by polarlipids. The intention of the review is to discuss the diverse nature of(polar) lipid–protein interactions and how it can affect the physicochemicalproperties of lipid–protein systems. We will mainly focus on aspects rele-vant for the properties of emulsions and foams. However, it must bepointed out that lipid–protein interactions at interfaces are of fundamentalimportance also in many biological processes. One such example is thepulmonary surfactants system, consisting of a complex mixture of proteinsand lipids, which is essential for the function of the lungs (30–32).Obviously, the function of the lungs involves the formation of lipid–pro-tein film at the air–aqueous interface. Recently, the alveolar surface hasbeen suggested to be lined by a liquid-crystalline phase (33–35), ratherthan a monolayer, which so far has been the traditional view as presentedby Clements in the 1950s (36). Lipase-catalyzed lipolyses is another inter-facial process that is of large physiological importance (37). In addition,lipase-catalyzed lipolyses has come to large industrial importance in manytechnical applications, including detergency and food processing (38).Because the lipases are water soluble and most of the natural lipidshave low aqueous solubility, lipolyses takes place at the aqueous–lipidinterface and therefore one often speaks of ‘‘interfacial activation’’ inconnection with lipase activity (39,40).


The structure of lipids in aqueous systems is discussed in Chapters 3and 4. General information of proteins and protein conformation can befound in most textbooks on biochemistry. There are also publications dedi-cated to protein structure, like the excellent book by Creighton (41) or quitea number of good reviews (42–45). Here, we will highlight only the struc-tural and physicochemical properties of importance for the understandingof polar lipid–protein interactions.

II. LIPIDS

Lipids can be divided into two major groups: polar and nonpolar lipids.The nonpolar lipids, primarily the triglycerides, have small polar groupsand, hence, show only limited interaction with aqueous systems. The polarlipids, however, with large charged or uncharged polar groups, givingthese lipids an amphiphilic nature, associate in aqueous systems. Thecommon feature for the self-assembly of the polar lipids in aqueous envi-ronment is the formation of a polar interface, which separates the hydro-carbon and water regions. The hydrocarbon chains can exist either in afluid state, as in liquid-crystalline phases, or in a solid state, as in the lipidgel phases (46). Generally, the melting of the chains in an aqueous envi-ronment occurs at a much lower temperature compared to the melting ofthe pure lipid.

Polar lipids can be divided further into two classes on the basis of theirinteraction with water:

1. Lipids and synthetic analogs that are water soluble in monomericand micellar form (i.e., surfactants)

2. Lipids with very low water solubility, but with the ability to swellinto liquid-crystalline phases

The water-soluble polar lipids (e.g., ionized fatty acids, bile salts, andsynthetic surfactants, charged or uncharged) have monomeric solubilityin the millimolar range and form micelles at higher concentrations. Thecritical micelle concentration (cmc) is considered to be a narrow concen-tration range, within which aggregates start to form by a strong coopera-tive process (47). The driving force for micelle formation is thehydrophobic interaction (cf. Ref. 48). The cmc for single-chain amphi-philes decreases with increasing chain length; for ionic amphiphiles, cmcalso depends on the ionic strength, because the addition of salt reduces theelectrostatic repulsion between the charged head groups. Increased tem-perature has, however, only a moderate influence on cmc, once the tem-perature has exceeded the critical temperature, where the monomer


solubility is equal to the cmc (Krafft temperature). At low water content,an inverse micellar structure, the L2 phase, is formed, in which the hydro-carbon chains form the continuous medium and the aqueous medium ispresent within the micelles.

A common feature of the two classes of polar lipids is the tendency toform lyotropic liquid-crystalline phases. A summary of some of the differentliquid-crystalline phases that can occur is given in Fig. 1. With decreasingwater content, the phase behavior of polar lipids often follows the sequencehexagonal phase (HI) ! lamellar phase (La) for water soluble lipids and

Figure 1 Commonly formed association structures by polar lipids. Phase tran-

sitions can be induced by changes in water content, temperature, or by interaction

with other solution components, like proteins. The lamellar liquid-crystalline phase

(La) can be regarded as the mirror plane, where the aggregates are of the ‘‘oil-in-

water’’ type on the water-rich side and of ‘‘water-in-oil’’ type on the water-poor side

(49). On both, the water-rich and water-poor sides of the La, there are two possible

locations for cubic phases. Other ‘‘intermediate phases’’ may also occur. The

formation of a particular phase can, in many cases, be understood by looking at the

geometric packing properties of the amphiphilic molecule in the particular envi-

ronment (50,51). This property can be expressed by the so-called packing parameter

(v/al), which is defined as the ratio between the volume of the hydrophobic chain (v)

and the product of the head group area (a) and the chain length (l ).


lamellar phase (La) ! reversed hexagonal phase (HII) for lipids with lowwater solubility. Cubic liquid-crystalline phases (Q) often occur in betweenthese. Phase transitions can also occur with changes in temperature; withincreasing temperature, the sequence of thermal transitions is usually thesame as with decreased water content. In many cases, the formation of aparticular phase can be understood by looking at the geometric packingproperties of the amphiphilic molecule in the particular environment(50,51), which is the cross-section area of the polar head group in relationto that of the acyl chain. This property can be expressed by the so-calledpacking parameter (v/al), which is defined as the ratio between the volumeof the hydrophobic chain (v) and the product of the head group area (a) andthe chain length (l ). The packing parameter for a particular environmentwill determine the curvature of the interface and, thus, the particular phase.Generally speaking (see Fig. 1), a value of the packing parameter lower thanunity (cone-shaped amphiphile) facilitates the formation of structures wherethe polar interface is curved toward the hydrocarbon phase (i.e., structuresof ‘‘oil-in-water’’ type (L1, HI). On the other hand, a value larger than unity(reversed cone-shaped amphiphile) will give the reverse curvature and favor‘‘water-in-oil’’ structures like HII and L2. When the packing parameter ischanged, for instance, by the change of ionic strength or temperature or theaddition of other molecules like proteins, phase transitions will ultimatelyarise. Increased temperature, for example, will increase chain mobility andthereby increase the volume of the lipophilic part of the molecules, explain-ing the often seen thermally induced transition La ! HII. Decreased hydra-tion will decrease the head-group repulsion, resulting in a decreasedinterface area and, thus, in an increase of the packing parameter.

In nature and in many technical applications, the lipid aggregatesconsist of a mixture of different lipids, which either exist in a homogenousmixture or separate into domains. As discussed in the review by Raudino(52), the lateral distribution in these mixed aggregates is influenced by anumber of factors like ionic strength, presence of polymers/proteins, as wellas the composition of the lipids, and it is thus hard to give any general rulesto predict when phase separation will occur.

Luzzati and co-workers determined the main features of the mostcommonly found mesophases in the early 1960s by x-ray diffraction[reviewed by Luzzati in 1968 (53); see also Chapter 3 of this book].

Results from spectroscopy studies have increased the understandingof the dynamic nature of these phases. The lamellar phase (La) consists ofstacked infinite lipid bilayers separated by water layers, whereas the hex-agonal phases consists of infinite cylinders, having either a hydrocarboncore (HI) or a water core (HII). As shown in Fig. 1, the cubic phases (Q)can exist in several locations in the phase diagram. The cubic phases have


been shown to exist in a number of lipid systems (54,55). They are iso-tropic and highly viscoelastic. Different structures of the cubic phases,depending on the particular lipid system, have been suggested (49,55–59).One type of cubic structure, consisting of rodlike aggregates connectedthree and three at each end, has been proposed by Luzzati et al. (56).The two polar rodlike networks formed in this way gives two continuousand unconnected systems of fluid hydrocarbon chains. We will, however,mainly focus on the other main type of cubic phase, which has beenobserved in aqueous dispersions of polar lipids with low aqueous solubilitylike monoglycerides, phospholipids, and glyceroglucolipids (46,54,57), aswell as for water-soluble surfactants like ethoxylated fatty alcohols (60).This type of cubic phase is bicontinuous and based on curved noninter-secting lipid bilayers (cf. Fig. 10) that are organized to form two uncon-nected continuous systems of water channels (cf. Refs. 55,57, and 61 andChapter 11 of this book). The two principal radii of curvature, R1 and R2,can be used to describe the curvature of any surface. The average curva-ture 1=2ð1=R1 þ 1=R2Þ is by definition zero at any point for a minimalsurface. Thus, at all points, the surface is as concave as it is convex. If aninterface is placed in the gap between the methyl end groups of the lipid inthe bicontinuous bilayer type of cubic phase, it will form a plane that canbe described as a minimal surface (57,62). A minimal surface exhibitingperiodicity, like in the cubic lipid–aqueous phase, is termed an infiniteperiodic minimal surface (IPMS). It has been shown by Hyde that thepacking parameter for a lipid in such a curved bilayer is larger thanunity and can be related to the Gaussian curvature [1/(R1 R2)] of theIPMS. Three types of IPMS, described by different cubic space groups,have been shown to be important in lipid systems (57,59,62):

. The primitive lattice (Pn3m) which corresponds to the diamond (D)type of IPMS

. The body-centered lattice (Ia3d) which corresponds to the gyroid(G) type of IPMS

. The body-centered lattice (Im3m) which corresponds to theprimitive (P) type of IPMS.

The occurance of micellar cubic phase, Cmic, space group Fd3m, wheredisjointed reversed micelles embedded in a three-dimensional hydrocarbonmatrix are organized in a cubic symmetry has been reborted by Luzzati andco-workers (64). The formation of this type of Cmic phase has previouslybeen reported for aqueous systems containing monoolein and oleic acid(64–67), for aqueous mixtures of sodium oleate and oleic acid (68), and,consequently, also during lipase-catalysed lipolysis of monoolein in aqueousdispersions under neutral/alkaline conditions (69,70).


Today cubic lipid–aqueous phases are recognized as important in bio-logical systems (34,46,55,57,58,59,66,71–75). Some of these reports suggestthat cubic lipid–aqueous phases can occur during the fusion of biologicalmembranes. There are vast amount of studies of membrane fusion [cf. thecomprehensive reviews by Kinnunen and Holopainen (2)], which isimpossible to cover here. It is however worth mentioning that early studieson the fusion process for biological membranes reported that lipidic particleswere discovered during the fusion event (76), indicating that additional liquid-crystalline phases occur during the fusion process. The nature of theseintermediate phases during fusion is unclear, although the formation ofinterbilayer structures as inverted micelles have been proposed (77). Inaddition, the fusion rate is increased in systems which have a lamellarphase! inverted hexagonal phase transition (76,78,79). As will be discussedfurther, such a phase transition can be affected by proteins and polypeptidesand there are also several reports on the promotion of fusion by proteins andpolypeptides (52).

The liquid-crystalline lipid–aqueous phases can exist in excess of aque-ous solution. One example of such lipid dispersions is unilamellar ormultilamellar vesicles,* which are formed from lamellar (La), phases. Thestability, size, and shape of vesicles can vary, depending on the compositionof lipids and aqueous phase (for reviews, see Refs. 80–83). In analogy withliposomes, dispersions of a cubic lipid–aqueous phases, cubosomes, whichwere first discovered by Larsson et al. (57,75,84) are also formed with anexcess of water. The stability of cubosomes, formed in MO–H2O-basedsystems, and the corresponding dispersed HII phase (hexosomes)in the MO–TO–H2O system was found to increase in the presence of anamphiphilic block copolymer (polyoxamer) (84–86).

III. PROTEINS

An important consequence of protein–lipid interaction is the effect on thestability of the protein in solution as well as its behavior at interfaces. Whendiscussing the stability of proteins, we may distinguish between the confor-mational stability of proteins and aggregation/precipitation phenomena dueto reduced solubility at pH close to the isoelectric point, at high ionicstrength (salting out), and/or caused by specific binding of ions (e.g., theformation of calcium bridges) or lipids. Although the two phenomena

*The term liposomes are, according to IUPAC recommendation, synonymous to lipid vesicles,

but are sometimes used for multilamellar vesicles.


usually are connected, aggregation/precipitation can occur without majorconformational changes of the protein (87). The conformational stability ofa protein, which, of course, has no meaning for proteins lacking secondarystructure, can be estimated by circular dichroism (cf. Ref. 41), compres-sibility measurement (cf. Ref. 88), and calorimetry (cf. Refs. 89–91). Thestabilization of the protein structure have been extensively reviewed by anumber of authors (cf. Refs. 43,44, and 89–92), and we will only focus onsome aspects of significance in emulsion systems.

The native protein structure is a consequence of a delicate balance offorces, including electrostatic forces, hydrogen-bonding, van der Waalsforces, conformational entropy, and so-called hydrophobic interactions(cf. Refs. 43 and 91–94). The amino acid sequence of the polypeptidechain (the primary structure) will determine the folding into structuralunits (the secondary structure), and the association of structural units intodomains, tertiary and quaternary structures, gives each protein theunique conformation connected with its action and specificity. Naturally,cross-links, such as disulfide bridges, increase the stability of a protein.

The interior of a globular protein is very densely packed, having aquite constant mean packing density (0.74), a value also found for crystalsof small organic molecules (94). Thus, van der Waals forces and hydrogen-bonding, which are short-range interactions, play an important role for thestability of folded proteins (91). As first pointed out by Kauzmann (95), it isclear that the so-called hydrophobic interactions play an important role instabilizing the protein structure. The nonpolar amino acid residues willprovide a strong driving force for folding, leading to an accumulation ofhydrophobic residues in the core of the protein molecule. The polar aminoacid residues (uncharged and charged) will have a high affinity for aqueoussolvent and will, consequently, be located on the outside of the protein. Thenature of hydrophobic interactions in this context is not yet fully under-stood (cf. Refs. 43,91, and 92), because it still is difficult to analyze themseparately from other forces contributing to the stabilization of the proteinstructure (91).

It is important to bear in mind that proteins are only marginally stableat room temperature. This means that the exchange of only one amino acidresidue, by, for instance, genetic engineering, might destabilize or stabilizethe protein considerably. In the following, it will be demonstrated that thebinding of lipids to the protein also can have the same effects. In manysystems involving proteins and lipids, the protein–lipid interaction takesplace at an interface. As discussed extensively by Norde et al. (96–98), thedelicate balance between forces that stabilize and destabilize the proteinmight be shifted in the proximity of an interface, leading to unfoldingupon adsorption. The loss of entropy upon protein folding is the main


force counteracting the stabilization of the protein structure (43). Thus,unfolding upon adsorption is an entropically favored process (96–98).Furthermore, at an interface, the unfolded hydrophobic domains might beoriented in such a way that their exposure to the aqueous environment isminimized. In fact, Norde argues that the entropy gained by the unfoldingof the protein upon adsorption can be a significant driving force for adsorp-tion (96–98). However, they also observed that adsorption of protein onapolar surfaces might lead to an increase order of the protein secondarystructure as observed for enzymes like a-chymotrypsin and serine proteinasesavinase on Teflon (98–100).

The folding and unfolding of proteins have been shown, under cer-tain conditions, to occur via an intermediate state, the molten globule state(101–106). This state, which is somewhere between the native and comple-tely unfolded state, is characterized by a retained secondary structure, butwith a fluctuating tertiary structure. The protein molecule is also moreexpanded and exposes more hydrophobic domains. The molten globulestate is hard to detect by calorimetric measurements, because the unfoldingof the molten globule is accompanied with little or no heat absorption(104). As discussed by Dickinson and Matsumura (106), the molten glo-bule state can be achieved in a number of ways, as pH changes, increase oftemperature, the use of denaturation agents, breaking of disulfide bridges,and removal of ligands or cofactors bound to the protein. It is also sug-gested that the protein can adopt a molten globule state when interactingwith an interface. In fact, it was found that a-lactalbumin was more sur-face active under conditions where it exists in the molten globule state. Ithas been proposed that the molten globule state of the protein may berequired for the translocation of proteins across biological membranes(107,108).

It should be noted that although it is possible to classify proteins asglobular, unordered (lacking secondary structure), and fibrous, there aremarked differences between proteins belonging to the same class, due totheir complex and intriguing design. Thus, it is much harder for proteinsthan for polar lipids to follow rules of thumb for their behavior. Proteinproperties such as conformation, charge distribution, association, and activ-ity are also strongly influenced by environmental condition (e.g., pH, ionicstrength, type of ion, and temperature). In this context, it is important topoint out the effect of type, valence, and ionic strength of an added electro-lyte. This can have a profound effect on interactions involving proteins andother polyelectrolytes, in particular under physiologically relevant condi-tions as discussed by Ninham et al. (109–111). They argue that the presenttheory is not sufficient to distinguish between van der Waals interactionsand electrostatic interactions. In addition, as part of their biological role,


some proteins have specific binding sites for lipids. These binding sitescan even be specific for a certain class of lipids. Thus, it is importantto consider protein–lipid interactions in relation to the features of eachindividual protein.

IV. INTERACTION BETWEEN GLOBULAR PROTEINS ANDWATER-SOLUBLE POLAR LIPIDS

A. In Solution

Ionic surfactants interact with most proteins. A high surfactant concentra-tion will generally lead to unfolding of the protein structure. The interactionbetween nonionic surfactants are weaker and seldom affect the structure ofproteins. Several reviews concerning the interaction between water-solublepolar lipids and protein are focused on the interaction between ionic surfac-tants [e.g., sodium dodecyl sulfate (SDS) and globular proteins at low andintermediate temperatures] (11,112–118). Because vast amounts of the sur-factant–protein work is devoted SDS, we will use this system as an example,and at the end of this subsection, we will discuss some exceptions.

We can distinguish between two types of binding of water-solublelipids to proteins:

1. A high-affinity type of binding that occurs at low lipidconcentration

2. Nonspecific cooperative interaction taking place at higherconcentrations (115,116).

An example of a binding isotherm, where the two types of bindingoccur, is given in Fig. 2. In this isotherm, for the binding of SDS to lyso-zyme, the regions of high-affinity noncooperative binding, at low surfactantconcentration, is well separated from the cooperative binding observed at ahigher concentration. For comparison, an example of a binding isotherm forthe binding of a nonionic surfactant, n-octyl-b-glucoside, to the sameprotein is also inserted. In this case, only nonspecific cooperative bindingoccurs.

The specific binding is mediated by ionic and hydrophobic interactionsand usually occurs below the cmc of the surfactant (115,119–122).

The nonspecific interaction occurs close to or above the cmc andusually leads to a destabilization of the native conformation. The cmc ofthe surfactant is thus an important parameter and conditions that affect cmcwill generally affect the binding (cf. Refs. 116 and 123). The saturation of allof the binding sites generally corresponds to 1–2 g of surfactant per gram ofprotein (115,116,124). The extent of interaction and unfolding depend


mainly on the nature of the surfactant hydrophilic group, surfactant chainlength, ionic strength, pH, temperature, and organic additives as well as onthe protein itself (116). Organic additives include the presence of impuritiesin proteins as well as in the lipids. For instance, it has been demonstrated byLunkenheimer and co-workers that commercial SDS samples usually con-tains a substantial amount of decanol, which actually is more surface activethan SDS in itself (125–127). Similarly, it has been shown by Clark et al.that b-lactoglobulin contain bound fatty acids, which may alter thebinding of other surface-active compounds (128). Clearly, the presence ofamphiphilic impurities may give anomalous effects on the binding of othersurfactants.

Figure 2 Binding isotherms for binding of surfactants to lysozyme in aqueous

solution at 25�C. The isotherms (s, f) for SDS have regions of both high-affinity

noncooperative binding, at low surfactant concentration, and cooperative binding at

high concentration. The influence of ionic strength on the binding isotherm is shown:

s, ionic strength (I )¼ 0.0119M and f, ionic strength¼ 0.2119M at pH 3.2. For

comparison, an example of a binding isotherm where only nonspecific cooperative

binding occurs is also inserted. This isotherm, describing the binding of the nonionic

n-octyl-b-glucoside (OG) to lysozyme (œ) was measured at pH 6.4 and ionic strength

of 0.132M. The protein concentration was 0.13% (w/v). The arrows indicate the cmc

for the different surfactants and ionic strengths. The data are adopted from Ref. 115

and the experimental details are given in Refs. 160 and 115 for SDS and OG,

respectively.


1. Anionic

The effect of surfactant protein interaction on the structural stability ofproteins depends strongly on the mode of interaction. In fact, as shown inFig. 3, the same surfactant can act to both stabilize and destabilize, depend-ing on surfactant concentration as well as other solution conditions. At lowsurfactant-to-protein ratios, a high-affinity interaction between certainproteins and surfactants occurs. This interaction stabilizes the proteinstructure against thermally induced unfolding; thus, the thermally inducedtransition is shifted toward a higher temperature, as illustrated in Fig. 3 andpreviously reported by Hegg (129) for SDS and b-lactoglobulin. Similarfindings have also been reported for other protein–surfactant complexessuch as between fatty acids or SDS and bovine serum albumin (130) aswell as between palmitic acid and b-lactoglobulin (131). As discussed earlier,

Figure 3 The thermograms from top to bottom shows the thermally induced

unfolding of b-lactoglobulin (1.4mM in 60mM NaCl, pH 6) when increasing the

protein/SDS molar ratio. The cmc of SDS is 0.47mM at 25�C and �1mM at 90�C,

when taking into account the ionic strength of the protein solution. Assuming that

one SDS molecule is bound per b-lactoglobulin monomer, 3mM SDS has to be

added to reach the cmc of the surfactant at 90�C. The data are adopted from

Ref. 123, where the experimental details are given also.


increasing the free-surfactant concentration to the cmc gives rise to non-specific cooperative binding, which, in turn, can lead to unfolding of theprotein, as illustrated in Fig. 3 (123). This is in agreement with earlierreports, in which a surfactant ratio above 10mol of SDS per mole ofserum albumin or 1mol of SDS per mole of b-lactoglobulin monomerwas observed to cause unfolding of the protein (129,130).

Surfactant like alkylsulfates or alkylethersulfates interacting withproteins with opposite net charge (e.g., lysozyme or gelatine) might causeprecipitation of the protein–surfactant complex due to neutralization of thenet charge (119,132–137). Although the protein is precipitated, only smallchanges in the secondary structure occur. At an increased surfactant con-centration, the complex is dissolved and the protein starts to be unfolded.Generally, denaturation of proteins by long-chain alkyl sulfates such as SDSresults in a structure with large fractions of the polypeptide chain in ana-helical conformation (48,138,139). As a simple rule, proteins with a lowcontent of a-helix in their native form, such as concanavalin A, b-lactoglo-bulin, and ovalbumin, will increase in a-helix content upon interacting withSDS. The reverse is observed for proteins with a high a-helix content in theirnative form (e.g., myoglobin and serum albumin) (138). The structure result-ing from the interaction is thought to consist of helical segments with flex-ible joints and with most of the hydrophobic side chains exposed to thesurfactant. The successive binding of SDS opens up the molecules, due tothe increased electrostatic repulsion, and unveils new hydrophobic domains,which can bind additional surfactants. This association stabilizes a-helicalfolding at the expense of nonrepetetive structure. The free energy gained bythis process in most cases by far exceeds the unfavorable free-energy changeof disrupting the native conformation (48). Light-scattering studies confirmthe expansion of the hydrodynamic radius of the protein upon interactionwith SDS (140). Several models of the structure of complexes between SDSand proteins at high surfactant concentration, like the correlated necklace,rodlike structure, and flexible helix, have been considered (cf. Refs. 116 and141). However, small-angle neutron-scattering data strongly indicates astructure resembling a necklace (141,142), where the polypeptide chainwith high flexibility is decorated with SDS micelles (138,141), as shown inFig. 4. This interaction is reported to take place via the monomeric form ofthe surfactant (116,138).

It should also be kept in mind that not all proteins are fully unfoldedby SDS. For instance, it has been shown that the activities of glucoseoxidase, papain, pepsin, and bacterial catalase were not affected by highconcentration of SDS, correlated to the low binding of SDS (143,144).

Thus far, we have mainly addressed the interaction at low proteinconcentrations. Moren and Khan (136) investigated the phase behavior of


the anionic SDS, positively charged lysozyme, and water over a wide con-centration range and one of the phase diagrams they determined is given inFig. 5a. Stenstam et al. later investigated in detail the stoichemistry of theformed complex and their findings are summarized in Fig. 5b (137). Smallamounts of SDS, at a ratio to lysozyme corresponding to charge neutraliza-tion of the protein, were found to give precipitation. A net attractive forceexists between the surfactant–protein complexes and hydrophobic interac-tions dominate (Fig. 5b). Further addition of SDS lead to a redissolution ofthe precipitate, which was complete when the number of SDS molecules wasequal to the number of positive charges on the protein (19). At this ratiobetween SDS and lysozyme, a narrow strip of a bluish gel phase occurredwhen the protein concentration was between 7% and 20% (w/w). At ahigher ratio between SDS and lysozyme, the interaction between the surfac-tant protein complexes is net repulsive and electrostatic interactionsdominate (Fig. 5b). Consequently, an isotropic solution is formed. Morenand Khan also investigated the effect of varying alkyl chain length, C12SO4,C10SO4, C8SO4, and C6SO4 on the lysozyme–sodium alkyl sulfate–waterternary systems (145). The surfactant with shortest hydrophobic tail(C6SO4) forms the largest solution region with lysozyme prior to precipita-tion, and the extension of the solution region decreases with increasingsurfactant chain length. The extension of the precipitation region towardhigher surfactant concentrations increases with decreasing surfactant chainlength. The surfactant concentration required to redissolve the precipitate atdilute protein concentrations therefore seems to follow the cmc for thesurfactant in water, which also increases with decreasing surfactant chain

Figure 4 Schematic representation of the so-called necklace model for the

interaction between SDS and proteins. The solid line represents the unfolded

polypeptide chain, which still contains secondary structure. Micellelike clusters are

cooperatively formed on the polypeptide chain.


length. A single gel phase was only observed for the C12SO4 and C10SO4

systems and not in the presence of C8SO4 and C6SO4. A similar type ofgel phase is expected to occur in more food-relevant surfactant–lipid andprotein–aqueous ternary system and therefore offer potentially veryinteresting possibilities to vary the functional properties of foods and foodingredients.

Within the type of surfactant, the binding is dependent on the natureof the polar head group; for example, for an anionic surfactant, the

Figure 5 (a) Phase diagram of the lysozyme–SDS–water ternary system, where L

indicates solution, G gel, and P precipitate. (Adapted from Ref. 136, where

experimental details are given.) (b) Schematic representation of the interaction

between protein surfactant complexes in the lysozyme–SDS–water system. (Adapted

from Ref. 137, where the experimental details are given.)


interaction decreases in the order alkyl sulfates>alkyl sulfonates>alkylbenzene sulfonates>carboxylates � alcohols (146,147).

2. Nonionic

The interaction between nonionic surfactants and proteins is generally weak(118,146,148–151). In the case of ionic surfactants, the specific ionic inter-action, absent for a nonionic surfactant (Fig. 2), occurs in addition to thehydrophobic interaction, leading to more severe effects on the protein struc-ture. For instance, each b-lactoglobulin monomer binds only one Tween-20(152), or one sucrose ester (153), or one Triton X-100 (149). Generally,minor changes of the structure upon interaction are observed (148,150).An unordered, flexible protein, b-casein, was found to bind less than onesucrose ester per protein molecule, possibly due to incorporation of thesurfactant in b-casein micelles (153). The reason for the weaker interactionbetween proteins and nonionic surfactants has been assigned to the lowercmc, which gives a too low monomer concentration to attain cooperativebinding to the protein (148). The cmc is increased when the chain lengthis decreased, which may change this situation; The binding of octyl gluco-side to various proteins was found to occur in a cooperative manner atsurfactant/protein molar ratio of 100 and more, without any evidence ofprotein denaturation (150).

Also, the nature of the nonionic polar head groups will affect theinteraction. For a series of Triton X surfactants, increasing the hydrophilicoxyethylene chain length was found to decrease the strength of interactionwith bovine serum albumin (BSA), due to steric hindrance (151). The calori-metric data indicated that some conformational changes of BSA occurredduring the saturation of the low-affinity, noncooperative binding sites (151).

Some studies have also been carried out with the zwitterionic surfac-tant lysophosphatidylcholine (LPC), which was found to bind cooperativelyto puroindoline, a lipid-binding protein isolated from wheat flour, at amolar ratio of 5–1 (154). One LPC molecule was also found to bind withless affinity to b-lactoglobulin than Tween-20 (21). The binding of Tween-20, as opposed to LPC, had a much more disruptive effect on the interfacialfilm of the protein, attributed to the bulkier head group of Tween-20. Thisimplies that also a nonionic surfactant can disrupt the structure of a protein,provided that the binding is strong enough and the hydrophilic head grouplarge enough to sterically induce conformational changes.

3. Cationic

Cationic surfactants generally seem to exhibit an intermediate action onwater-soluble proteins. Reports in the literature indicate a cooperative


interaction with proteins, but with less affinity and, thus, with less perturba-tion of the folded state, compared to the effect of the anionic ones (123,155–159). If the binding is governed both by electrostatic and hydrophobicinteractions, anionic and cationic surfactants will obviously occupy differentsites. Nozaki et al. have suggested that the lower affinity of many proteins forcationic compared to anionic surfactants can be explained by the fact that thecationic arginine and lysine side chains contribute more CH2 groupsthan anionic aspartate and glutamate side chains (157). This implies thatthe combined electrostatic and hydrophobic interactions will be morefavorable for anionic surfactants. As a consequence, the cooperative bindingstep will start at a higher concentration for cationic relative to anionicsurfactants (116).

4. Effect of Solution Conditions

Increased ionic strength affects the interaction between protein and ionicsurfactants by reducing the electrostatic attraction between surfactants andamino acid residues with opposite net charges. Generally, the high-affinitynoncooperative binding is strongly influenced by the electrostatic inter-action between surfactant and protein. Thus, this part of the bindingisotherm will be shifted toward higher surfactant concentration upon theaddition of salt, as observed for lysozyme and SDS (Fig. 2) (115,160).Increasing the ionic strength will, on the other hand, favor the cooperativebinding by screening the repulsion between the charged surfactant headgroups. This part of the surfactant-binding isotherm will therefore beshifted toward lower surfactant concentrations, parallel to the decreaseof surfactant cmc. Here, it is important to point out that the presenceof highly charged proteins as a polyelectrolyte as well as temperaturewill affect the formation of micelles. This has been amply demonstratedby Waninge et al. who studied thermally induced unfolding of b-lactoglo-bulin at a concentration of 1.4mM in 60mM NaCl, pH 6, in the presenceof various molar ratio SDS, and their main findings are illustrated by thethermograms in Fig. 3 (123). From Fig. 3, we note that the peak corre-sponding to the thermal unfolding disappears when the protein/SDS molarratio increases above 1:2. This corresponds to a SDS concentration ofabout 3mM. The cmc for SDS is about 8.1–8.2mM in water (161,162).However, the cmc for ionic surfactants decreases with ionic strength andincreases with temperature (161–164). Taking account to these effects forthe presence of b-lactoglobulin, which has a net charge of –5, at a con-centration of 1.4mM in 60mM NaCl, the cmc of SDS is expected to be0.47mM at 25�C and �1mM at 90�C. When taking into the specificbinding of one SDS molecule per b-lactoglobulin monomer, 3mM SDS


has to be added to reach the cmc of the SDS at 90�C. Thus, any effect ofnonspecific cooperative interaction between the surfactant and the proteinis expected to take place at this SDS concentration. In Fig. 3, we observean apparent loss of protein structure. The unfolding of the protein struc-ture at low temperature, which is observed in the presence of most anionicsurfactants such as SDS at high concentration, is expected to be main-tained at increased temperature. However, because the cmc generallyincreases with temperature, we might arrive at the situation where thecooperative binding ceases to exist at the high temperature, maybe evenbelow the temperature at which thermally induced unfolding take place.Interestingly, Waninge et al. observed that the conformational changesinvoked by the nonspecific cooperative binding of SDS at 25�C could bereversed by extensive dialysis (123). Referring to the discussion of themolten globule state earlier, it is tempting to compare it with the structureobtained by surfactant unfolding.

Although cationic surfactants seem to cause less unfolding of globularproteins at low temperature than anionic, some reports indicate that theycan destabilize globular proteins at increased temperature (158,165).However, these reports also indicate that the unfolding process at thesame time becomes considerably more reversible. The heat denaturationof ovalbumin, which in practice is completely irreversible, was found tobe completely reversible in the presence of high concentrations of cationicsurfactants (165). This was explained by decreased intermolecular and intra-molecular interactions at high temperature, due to the interaction betweenthe unfolded protein and surfactant, which facilitates the reformation of thenative complex on cooling.

As a rule of thumb, an increase in pH will shift the binding of anionicsurfactants to higher concentrations (166). A decrease of pH will have thesame effect on the binding of cationic surfactants (167). At low surfactantconcentrations (i.e., well below the cmc), cationic amphiphiles increase thesolubility of proteins on the acidic side of the isoelectric point (pI ), whereasprecipitation can occur on the alkaline side of the pI. Anionic amphiphileswill affect solubility in the opposite direction. The solubilising effect is alsoobserved at high temperatures.

We conclude that because the binding generally is thought to occur viamonomers, any change affecting the cmc will also affect the cooperativebinding at concentrations close to and above the cmc. Under someconditions, the formation of surfactant micelles will be energeticallyfavored before binding to the protein. If the cmc is of the same order ofmagnitude as the concentration necessary for binding to occur, thelowering of the cmc caused by increasing ionic strength might even preventbinding.


B. At Interfaces

The stabilization of food emulsions and foams can be achieved by proteins,polar lipids, or mixtures of them. However, the mechanisms by which theystabilize emulsions and foams can be quite different. Here, we refer to thework of Clark and his colleagues (168), which have, based on results such asthose from fluorescence recovery after photobleaching experiments, sug-gested a mechanism for stabilization of foams and emulsions. Generally,polar lipids are capable of reducing the interfacial tension more than pro-teins, whereas the protein molecules can be anchored at multiple sites at theinterface. Therefore, many proteins stabilize foams and emulsions by form-ing intermolecular interactions between the adsorbed protein molecules,which encapsulates the dispersed phase. Such a mechanically rigid layerimmobilizes the proteins and prevents perturbations, droplet coalescence,or flocculation. In contrast, lipids usually stabilize the dispersed droplet orbubble by forming a densely packed and much less rigid monolayer with highmobility. Any disturbance of the interface, such as expansion of the interface,will induce a (temporary) surface-tension gradient. This is, of course, unfa-vorable and the system tries to restore an even distribution of the surface freeenergy. During the lateral diffusion, or adsorption from the bulk, of thesurface-active components, a flow counteracts the disturbance. This dynamicprocess is usually called the Gibbs–Marangoni effect (169). Thus, the rate ofexchange of polar lipids at the interface is much higher than for proteins. Theaddition of polar lipids/surfactants to a protein film can, if the concentrationbecomes too high, destroy the integrity of the film by disrupting protein–protein interactions and, consequently, destabilize the foam or the emulsion.This points at another important reason to study the interaction betweenproteins and polar lipids in the adsorbed layer as well as during the compe-titive adsorption of the individual components. The surfactant-to-proteinratios as well as the properties of the components and the interface willdetermine the composition of the film.

In several investigations, surface-tension measurements have been usedto study protein–lipid interactions (cf. Refs. 10,12,13,170, and 171). How-ever, it must be kept in mind that any impurity with a higher surface activitythan the studied components will accumulate at the interface, giving a low-ering of the surface tension (125–127) and thus affecting the interpretation ofthe data. The presence of impurities (e.g., fatty acids) bound to b-lactoglo-bulin did have a profound effect on the interfacial behavior of mixtures withTween-20, as judged from surface elasticity measurements at the air–aqueousinterface (128). It was observed that the film containing purified b-lactoglo-bulin could maintain a more rigid film, at a much higher concentration ofTween-20 as compared to the sample containing impurities.


We will discuss some experimental data on the basis of possible factorsthat influence the way mixtures of proteins and polar lipids behave at aninterface. These can be summarized as follows:

1. The surface activity of the individual components

(a) Competitive adsorption. The lipids and proteins compete forthe interface, where the most surface-active and/or abundantmolecule wins, depending on the ratio between surfactantsand proteins in solution.

(b) Displacement. The polar lipids may, due to their highersurface activity, displace the proteins from the interface. Thisdisplacement can be hampered by a strong interactionbetween the protein and the interface and/or protein–proteininteractions.

2. Protein–lipid interactions

(a) Increased surface activity of the lipid–protein complex

(i) The binding will cause unfolding and/or increase hydro-phobicity of the protein that will lead to an increasedaffinity to the surface.

(ii) The binding (of ionic amphiphiles) will cause precipita-tion at the interface due to charge neutralization.

(b) Decreased surface activity of the lipid–protein complex

(i) The binding will make the protein more soluble andhence lower the affinity for the interface.

(ii) The binding will lead to precipitation of protein–lipidcomplex in the bulk, which will cause loss of surface-active material.

(c) Protein–lipid interactions at the interface

(i) The interaction will give more efficient packing at theinterface and thus give a higher total surface concen-tration.

(ii) The interaction will disrupt the protein–protein inter-action in the interfacial film.

We will use some of our surface-tension data on protein–surfactantmixtures to demonstrate the effect of different modes of interaction. Differentmodes of interaction are observed for the same system depending on thelipid/protein ratio. The data will also serve as the basis for the discussionaccording to the above factors. The interaction between a protein


(ovalbumin) and three types of surfactant [an anionic (SDS), a nonionic(1-monocaproin) and a cationic (hexadecylpyridinium chloride)] as observedfrom surface-tension isotherms is illustrated in Fig. 6 [the experimental pro-cedure is given elsewhere (171)]. The protein concentration used was 21 mM.To avoid precipitation, the measurements were carried out at pH 5.6 and 4.0for SDS and hexadecylpyridinium chloride, respectively [i.e., below andabove the isoelectric point (4.5–4.9) of ovalbumin]. The interaction betweenovalbumin and ionic surfactants is considered to be nonspecific. Serum albu-min has, however, specific binding sites for the surfactants (114) and, forcomparison, the data for the surface tension of BSA (13 mM, 0.05M phos-phate buffer, pH 5.6) and SDS are also inserted in Fig. 6b. The surface-tension isotherms of the pure surfactants are shifted according to the differ-ences in cmc although they have a similar shape. Hence, it is useful to discussdata in terms of the ratio between added number of surfactant monomersand protein molecules.

1. Competitive Adsorption

No synergistic effect is observed for the mixture of the nonionic surfactant,monocaproin, and ovalbumin (Fig. 6a). The component giving the lowestsurface tension displaces the other at the interface. No molecular interac-tions seem to take place in either the bulk solution or at the interface. Also,for the cationic and anionic surfactants, the surface tension at high enoughsurfactant concentration is dominated by the contribution from surfactants.Thus, it is generally observed for a range of surfactant–protein systems thatthe protein dominates at low surfactant concentration, whereas the surfac-tant dominates at high surfactant concentration, as they generally give alower surface tension [13].

The effect of interfacial properties of a solid surface on the competitiveadsorption among a protein, fibrinogen, and a nonionic surfactant, penta-ethyleneglycol mono n-dodecyl ether (C12E5), was investigated by using awettability gradient silica surface (172). Fibrinogen was preferentiallyadsorbed on the hydrophilic surface at all concentrations, whereas thesurfactant dominated at the hydrophobic and intermediate part ofthe gradient at surfactant concentrations close to the cmc and above. Thepreferential adsorption of the surfactant extended to a wider range of thegradient if the temperature was increased to a value close to the cloud pointof the surfactant.

2. Displacement

Mackie et al. have, based on extensive studies using atomic force micro-scopy (AFM), Brewster-angle microscopy (BAM), fluorescence microscopy,


Figure 6 (a) Surface-tension isotherms of 21 mM ovalbumin (OA) (œ) in the

presence of the nonionic monocaproin (MC) in water adjusted to pH 5.6, where the

surface tension of the pure protein is marked with an arrow on the ordinate. Surface

tension of pure MC is also shown (s) and the cmc is marked with an arrow on the

abscissa. The surface-tension measurements were performed according to the drop-

volume method as a function of time. The surface-tension value after 2000 s has been

used for the isotherms. Further details are given elsewhere (171). (b) Surface-tension

isotherms of 21mM ovalbumin (OA) (œ) and 13 mM BSA (^) in the presence of the

anionic SDS in 0.05M phosphate buffer, pH 5.6. The surface tension of the pure

proteins are marked with arrows on the ordinate. The surface tension of pure SDS is

also shown (s) and the cmc is marked with an arrow on the abscissa. Other

conditions are the same as given for part (a). (c) Surface-tension isotherms of 21 mMovalbumin (OA) (œ) in the presence of the cationic hexadecylpyridinium chloride

(HPC) in water adjusted to pH 4.0. The surface tension of the pure protein is marked

with an arrow on the ordinate. The surface tension of pure HPC is also shown (s)

and the cmc is marked with an arrow on the abscissa. Other conditions are the same

as given in part (a).


and surface rheological techniques, developed an ‘‘orogenic’’ displacementmodel (173–175). This model is based on the adsorption of the displacingsurfactant/lipid into localized defects in the heterogeneous protein network.These nucleation sites then grows, which leads to a compression of theprotein network. During this initial stage, the protein film seems to bedenser but not thicker. However, above a certain critical density, the filmthickness starts to increase with the expanding surfactant domains whilemaintaining the protein film volume. Eventually, a protein networkcannot withstand the high surface pressure, but it will collapse and theprotein will be released and desorb from the interface. This model pointsat the importance of protein–protein and protein–interface interactions inrelation to the interaction of the surfactant with defects in the protein net-work as well as with the interface.

Indirectly, the neutron reflectivity study on the binding of SDS ontopreadsorbed layers of BSA at the hydrophilic silicon oxide–water interfaceby Lu et al. confirm the ‘‘orogenic’’ displacement model (176). Their resultsshow a uniform layer distribution of SDS at low surfactant concentrations,whereas the distributions become unsymmetrical as the SDS concentrationincreases. The binding of SDS results in an expansion of the preadsorbedBSA layer from 35 A in the absence of SDS to some 80 A at 3� 10�4MSDS, which Lu et al. interpreted as a considerable structural deformation ofthe protein. They based this interpretation on the close agreement betweenthe volume ratio of SDS to BSA in the mixed layer of 0.45 and the literaturevalue for the binding of SDS onto denatured protein in the bulk reported byTanner et al. (140). The specular neutron reflection is sensitive to the densityprofile normal to the interface, but it does not give any lateral resolution.Therefore, the observations by Lu et al. (176) can also be explained by the‘‘orogenic’’ displacement model (173–175).

Figure 6 Continued.


The effect of the interfacial protein film age on the displacement of theprotein from the surface of emulsion drops by nonionic water-soluble sur-factants [Tween-20 and octaethylene glycol n-dodecyl ether (C12E8)] showedthat b-lactoglobulin is harder to replace the longer the residence time was(177,178). Apart from the possible conformational changes that occur duringthe adsorption process, which can hamper displacement, it has been reportedthat b-lactoglobulin might polymerize through disulfide exchange at the oil–water interface (179). Consequently, the displacement of b-casein, which is aflexible and unordered protein without sulfhydryl groups, did not depend onthe age of the film. Furthermore, it was observed that it was harder to replaceb-lactoglobulin from a emulsion prepared close to the pI of the protein thanat neutral pH, whereas the replacement from emulsions prepared at pH 3 waseasier and no time effects was observed. Mackie et al. also studied displace-ment of b-lactoglobulin and b-casein by Tween-20, but from the air–waterinterface (173). They also found that b-casein was more easily displaced (i.e.,b-lactoglobulin films breaks at higher surface pressures). Furthermore, stressinvoked by penetration of the surfactant was found to propagate homoge-nously through the b-casein film, which, in turn, resulted in the growth ofcircular surfactant domains at the interface. b-Lactoglobulin, on the otherhand, was found to form elastic (gellike) networks at the air–water interfaceand the penetration of the surfactant therefore resulted in the growth ofirregular (fractal) surfactant domains. Interestingly, Tween-20 preferentiallydisplaced b-casein before b-lactoglobulin from a mixed b-casein/b-lactoglo-bulin film at the air–water interface (174).

The effect of altering the protein conformational stability on the dis-placement of adsorbed protein layers by surfactants was also visualized atthe solid–liquid interface by McGuire et al. (180). They found that theremoval of wild-type and structural stability mutants of bacteriophage T4lysozyme from hydrophobic and hydrophilic silica surfaces by a cationicdetergent, decyltrimethylammonium bromide (DTAB), generally increasedwith the stability of the mutants.

Wahlgren and Arnebrant (181) investigated the effect of the surfaceproperties on the displacement of adsorbed b-lactoglobulin (negative netcharge) and lysozyme (positive net charge) by the cationic surfactant cetyl-trimethyl ammonium bromide (CTAB) and the anionic SDS. They usedhydrophobic (hydrophobised silica), negative (hydrophilic silica), neutral(chromium oxide), as well as positively charged (nickel oxide) surfacesand found four types of behavior for surfactant concentrations well abovethe cmc:

1. Surfactant binds to the protein and the complex desorbs ondilution. This was observed for SDS and b-lactoglobulin as well as


lysozyme on a negative silica surface and can be explained bysimple electrostatic considerations. No adsorption from SDS–protein mixtures occurred.

2. The surfactant replaces the protein at the interfaces. This requiresthat the surfactant interacts more strongly with the surface thanthe protein, as was observed for CTAB with negative silica andSDS and CTAB with the hydrophobic surface when the adsorbedlayer consisted of b-lactoglobulin.

3. The surfactant adsorbs reversible on top of the protein layer. Theprotein–surface interaction is the stronger one and the surfactantis thus unable to solubilize the protein from the interface. This wasobserved for CTAB interacting with both proteins at thechromium oxide surface and SDS interacting with b-lactoglobulinat the nickel oxide surface.

4. Partial removal of the protein. This can be explained as due to thepresence of multiple binding sites for the protein, which will give acombination of mechanism 1 and 2.

Others have made similar observations. For instance, Green et al. alsostudied the interaction between SDS and preadsorbed lysozyme at thehydrophilic silicon oxide–water interface by neutron reflectivity measure-ments (182). SDS binds cooperatively to the preadsorbed protein layer atintermediate surfactant concentrations, with no desorption of lysozymefrom the interface. The protein was partly removed when the SDS concen-tration was increased to above 0.5mM, whereas a surfactant concentrationof 2mM was required to completely remove both protein and surfactantfrom the interface in accordance with the type 1 behavior discussed.

3. Increased Surface Activity of the Lipid–Protein Complex

When comparing the data for the interaction between SDS and ovalbuminand the corresponding data for BSA, we clearly observe the different modeof interaction (Fig. 6b). The gradual decrease in surface tension withincreasing surfactant concentration observed for ovalbumin and SDS mix-tures can be explained by more efficient packing at the interface as discussedhere. In addition, it has been argued that the electrostatic interactionbetween surfactant and protein might increase the hydrophobicity of theprotein and, hence, its surface activity. The specific binding of SDS to BSAdoes not affect the surface tension until the concentration corresponding tosaturation of the high-affinity binding sites is reached (i.e., 9–10mol SDSper mole protein) (114), where a sharp decrease in surface tension isobserved. This arises probably from an increase in the free-monomer con-centration of SDS. The second plateau, indicating constant surfactant


monomer concentration, which is observed at increased surfactant concen-tration, is likely to be connected with saturation of the cooperative bindingsites. As the surfactant concentration further increases, the surface-tensionisotherms for the two protein–surfactant mixtures coincide. The secondplateau observed in surface tension isotherms for ovalbumin and hexade-cylpyridinium chloride (HPC) mixtures just below cmc of HPC (Fig. 6c) canbe related to the electrostatic interaction between HPC and globular pro-teins that has been observed below the cmc in bulk solution (158). It isnoteworthy that the surface tension is slightly lower than for pure HPC,suggesting that the complex is more surface active. Green et al. used spec-ular neutron reflection and surface-tension measurements to study theadsorption of lysozyme and SDS at the air–water interface (183). Theirresults show that the lysozyme–SDS complexes are much more surfaceactive than the unbound species as the surface excesses for both lysozymeand SDS increases and surface tension decreases upon addition of SDS(region A). Interestingly, the molar ratio of SDS to lysozyme was foundto remain constant at about 7, although the total surface excesses increasewith SDS concentration up to a surfactant concentration of 2.5� 10�4M.This indicates that the complex that adsorbed on the interface had a ratherwell-defined stoichiometric composition.

A further increase in SDS concentration beyond 2.5� 10�4M led to asharp decrease in the total surface excess, whereas the molar ratio of SDS tolysozyme increased. Eventually, as more SDS was added, the mixed protein–surfactant layer was replaced by a pure SDS monolayer.

The surface activity of the complex depends also on the properties ofthe interface, as shown by Wilde and Clark (152) for liquid interfaces.They found that the complex between Tween-20 and b-lactoglobulin wasmore surface active at the oil–water interface than at the air–water inter-face, where the same surface activity as for the nonbound protein wasobserved. The complexes adsorbed at both type of interfaces was, how-ever, displaced by Tween-20 at the same surfactant-to-protein ratio. Here,we need to emphasize the difference in nature between the two types ofliquid interface, the liquid–air and the one between two condensed media,which explains the experimental observations. The oil–water interfaceallows hydrophobic residues to become dissolved in and interact favorablywith the oil phase, which is not possible at the air–water interface. Wehave also previously discussed that the unfolding of protein induced by theaction of surfactants or by the presence of an interface generally leads toexposure of hydrophobic residues; that is, the unfolded protein can besubstantially more ‘‘oil soluble’’ than the native one. This relates to thefollowing section, dealing with molecular interactions, where it will bedemonstrated that changes in oil-phase composition and hence solvent


properties also can lead to changes in the structure of the adsorbed proteinfilm. At the hydrophobized silica–aqueous interface, it was found that thecomplex between b-lactoglobulin and SDS, formed at a high surfactant–protein ratio, but still below cmc of SDS, was more surface active thanpure b-lactoglobulin (184).

4. Decreased Surface Activity of the Lipid–Protein Complex

The maxima in the surface-tension isotherm at HPC concentrations between0.8 and 2.5mM probably reflects an increased HPC–ovalbumin interactionin bulk solution (Fig. 6c). The formed highly charged complex is less surfaceactive and an increase in surface tension is thus observed. The surface-ten-sion maximum has been found to depend on ovalbumin concentration andis shifted toward higher HPC concentration at increased ovalbumin concen-tration (corresponds to 30mol HPC per mole ovalbumin, independent ofprotein concentration) (185). The adsorption from mixtures of humanserum albumin (HSA) and nonionic surfactant, decyl-dimethyl-phosphine-oxide (C10DMPO) at the air-water interface was reported by Miller et al.(186). They reported an anomalous surface-tension increase for the mixturesat low surfactant concentrations to values higher than for the protein at thesame concentration without the surfactant. Thus, it seemed that the surfac-tant–protein complex was less surface active. The likely explanation is thatthe nonionic surfactant is associated with HSA via hydrophobic interactionand thus makes the protein more hydrophilic and hence less surface active.Miller et al. also observed that the concentration range, where the coverageof protein and surfactant are comparable in the mixed surface layer wasquite narrow (186).

The precipitation of protein in the bulk solution due to neutralizationby added surfactant can also cause a decrease in surface concentration dueto loss of surface-active material. Garcia Dominguez et al. (187) have shownthat the surface-tension reduction of lysozyme and insulin at pH 3.5 (i.e.,below pI) decreased when an anionic surfactant (SDS) was added, due toprecipitation of the protein.

5. Protein–Lipid Interactions at the Interface

A synergistic effect on surface tension is seen for mixtures of proteins withboth the anionic and cationic surfactants (Fig. 6b and 6c). For ovalbuminand SDS mixtures (Fig. 6b), a gradual decrease of the surface tension withincreasing surfactant concentration is observed. This might be assigned tothe more efficient packing in the formed mixed surfactant–protein layercompared to the one formed by the individual components at this concen-tration (171). Even at the lowest concentration of cationic surfactant


(0.05mol HPC per mole ovalbumin), where the pure surfactant has the samesurface tension as water, a decrease in surface tension for the protein–surfactant mixture, compared to pure ovalbumin, is observed (Fig. 6c). Itis unlikely that any bulk interaction will affect the interfacial behavior at thislow HPC-to-ovalbumin ratio. Therefore, the lowering in the surface tensionprobably arises from molecular interactions in the adsorbed surface film,giving a more condensed surface layer. Buckingham et al. (188) found astrong synergistic lowering of the surface tension of a mixed solution ofSDS and poly-L-lysine at conditions at which no precipitation, micelle, orcomplex formation take place in the bulk solution. A similar behavior wasobserved for mixtures of low-molecular-weight surfactants of oppositecharges (189). This effect has been assigned to the formation of electro-neutral complexes in the interfacial film.

Not only the composition of the interfacial layer but also the mechan-ical properties (e.g., the dilational viscosity) of the layer are important forthe stability of emulsions and foams (8,11,14,190). In particular, both sur-face and bulk rheology as well as the disjoining pressure of the thin lamellaedetermine the stability of foams (14,191). Hence, in technical applications,thickeners are often added. The mechanical properties of interfacial filmscan, to a large extent, be controlled by the intermolecular interactions. Thismeans that the protein stabilization of a foam is mainly due to protein–protein interaction and the destabilization is thought of as a disruption ofthese interactions. Sarker et al. (21) discussed the effect of the surfactantproperties on the stability of interfacial films in foams. The addition of smallamount of lysophosphatidylcholine (LPC) was found to increase the foamstability of b-lactoglobulin foams (21). A further increase of the surfactantconcentration led to a decrease of the foam stability. The surface tensionversus molar ratio of LPC and b-lactoglobulin shows an inflection pointclose to a molar ratio of unity, corresponding to the binding of the surfac-tant to the protein. No increase of foam stability was, however, observed formixtures of Tween-20 and b-lactoglobulin, instead, the stability decreasedwith increasing surfactant concentration (192). The same observations weremade for the stability of an oil-in-water emulsion, where it was found thatsmall amount of Tween-20 increased the rate of shear-induced coalescenceof b-lactoglobulin-stabilized emulsion droplets (178). The marked reductionin surface shear viscosity even at low surfactant-to-protein ratios confirmedthat loosening of the protein layer occurred. The protein–surfactant com-plex is thought of being less surface active and a further increase of thesurfactant concentration will lead to replacement of protein and protein–surfactant complexes with the surfactant at the interface (192,193). Themobility of the protein in a protein-stabilized thin liquid film, as measuredwith the fluorescence recovery after photobleaching technique (FRAP),


increases at lower surfactant-to-protein ratio for Tween-20 than for LPC(Fig. 7). This was attributed to the stronger binding of Tween-20, comparedwith LPC, to b-lactoglobulin (21) and will also explain why the foambecomes unstable at a lower surfactant concentration when Tween-20 isused. The foaming properties of puroindoline from wheat was found to beimproved by the addition of LPC to protein molar ratio of 1–10 (154). Thiswas assigned to the forming of a complex, which prevents the interfacialaggregation of the protein. Once the surfactant concentration becomes largeenough, the protein–protein interactions within the surface film will beprevented, the mobility increased, and, thus, the foam stability decreased.

An ionic surfactant can also induce flocculation of protein-stabilizedemulsions and this is dependent on the nature of the protein–lipid inter-action, as discussed by Chen and Dickinson (134,135,194). An anionicsurfactant, sodium lauryl ether sulfate (SLES), at sufficient concentrationhas been found to flocculate gelatine-stabilized oil-in-water emulsion (134).A further increase in surfactant concentration was found to lead to arestabilization of the flocculated emulsion. In bulk solution, the anionicsurfactant will, at high enough concentrations, cause precipitation of thepositively charged gelatine. At a further increased surfactant concentra-tion, the precipitate was redispersed. Gelatine was initially displaced bySLES from the interface (194), but an increase of the surfactant concen-tration led to an increase of gelatine concentration at the interface and the

Figure 7 The effect of surfactant addition on the lateral diffusion in the adsorbed

mixed layer of surfactant and b-lactoglobulin, measured with the FRAP technique.

The diffusion coefficients of the fluorescent probe 5-N-(octadecanoyl)aminofluour-

escein and fluourescein isothiocyanate isomer 1 labeled b-lactoglobulin measured in

the presence of L-a-lysophosphatidylcholine (s) and Tween-20 (f), respectively, are

shown as a function of the molar ratio between surfactant and b-lactoglobulin. Thedata are adopted from the work of Sarker et al. (21) and Coke et al. (192),

respectively, in which the experimental details also are given.


surface charge became partly neutralized (135), causing flocculation. Afurther increase of the surfactant concentration led to a decrease of thegelatine surface concentration (194) and a restabilization of the emulsion(134). It was also observed that the addition of SLES to a b-lactoglobulin-stabilized emulsion not did cause any flocculation, although some kind ofcomplex was formed in bulk solution. It should be kept in mind thatb-lactoglobulin was negatively charged under the experimental conditionsused. This confirms the electrostatic nature of the observed SLES-inducedflocculation of the emulsions stabilized by the positively charged gelatine.Flocculation of b-lactoglobulin-stabilized emulsions was, however,observed in the presence of gelatine and SLES. Because it only occurredabove the cmc of the surfactant, it was suggested to depend on cross-linking of the emulsion droplets by surfactant micelles (134).

V. INTERACTIONS BETWEEN PROTEINS ANDWATER-INSOLUBLE LIPIDS

In this section, we will discuss interactions involving lipids with lowsolubility, where the lipids exist as dispersed particles, liposomes or vesicles,liquid crystalline phases, as well as monolayers at interfaces. Many of theprinciples discussed in the earlier sections also apply for protein–lipid inter-actions in condensed systems. Polar lipids, which normally are water inso-luble, associate into a variety of structures in aqueous solution. This processwill have an impact on interactions with proteins. For water-soluble surfac-tants, the association with proteins mainly involves a surfactant in themonomeric state, whereas for insoluble lipids, the association structuresalso have to be considered. We also note that even polar lipids that areconsidered water insoluble have a certain monomer solubility, which,although small (about 10�7 for mono-olein and about 10�10–10�12M forphospholipids), makes it possible for them to interact with proteins in themonomeric form—in particular, if the protein has a high-affinity bindingsite for the lipids. This is demonstrated in Fig. 8, which shows thethemograms from differential scanning calorimetry (DSC) measurementsof b-lactoglobulin, distearoylphosphatidic acid (DSPA), and b-lacto-globulinþ an aqueous dispersion of DSPA. We note that the peak corre-sponding to the thermally induced unfolding transition of b-lactoglobulin inthe presence of DSPA is shifted toward a higher temperature compared tothe one recorded for the pure protein. This confirms the presence of aspecific interaction between phosphatidic acid and b-lactoglobulin thatthermally stabilize the protein. This was also observed in the presence ofdipalmitoylphosphatidic acid (DPPA), but no such interaction was observed


when the protein was mixed with phosphatidylcholine, phosphatidyletha-nolamine, or phosphatidylglycerol (22). Neither could any interaction beobserved if the lipid contained unsaturated fatty acid residues. Thus, theresults show that the interactions between b-lactoglobulin and phospholi-pids are strongly dependent on the acyl chain as well as the head group. Asmall negatively charged head group is needed for the interaction to takeplace. Such an interaction can have important implication for the functionalproperties of the protein. We discussed earlier that fatty acids bound tob-lactoglobulin could affect the interfacial behavior of the protein (128).Kurihara and Katsuragi (277) reported that a lipid–protein complexformed between b-lactoglobulin and phosphatidic acid could mask abitter taste. This property was suggested to be specific for phosphatidicacid, as no effect was observed for mixtures of b-lactoglobulin andphosphatidylcholine, triacylglycerol, or diacylglycerol.

The large number of studies using lipid monolayers at the air–aqueousinterface and spread or adsorbed proteins have given us the basic knowledgeof the interaction between proteins and polar lipids with low aqueous solu-bility. Therefore, we will start to discuss some of the main conclusions thatcan be drawn from such studies. The following subsections will address theinteraction with oil–aqueous interface, vesicles, and liquid-crystallinephases.

Figure 8 The interaction between DSPA and b-lactoglobulin (b-Lg) is demon-

strated by the results from differential scanning calorimetry, where the thermogram

of the protein–lipid mixture is compared with those of the pure components.

Thermograms of DSPA, 5% (w/v) (—), b-Lg 5% (w/v) (– � –), and a mixture of b-Lg5% and DSPA 5% (w/v) (— - -—) in 1% sodium chloride at pH 7. A scanning rate

of 10�C/min was used. (Data from Ref. 22, where the experimental details are given.)


A. Lipid Monolayers at the Air–Aqueous Interface

1. Driving Force for Lipid–Protein Interactions

Electrostatics. Phospholipid–b-lactoglobulin interactions at the air–

aqueous interface has been investigated by Bos and Nylander (195) using thesurface film balance. Some of their findings are summarized in Fig. 9, where

Figure 9 The rate of incorporation of b-lactoglobulin into monolayers of DSPA,

DSPC, and DPPA versus surface pressure (�). The data were recorded at constant

surface pressure by measuring the area increase of the lipid monolayer spread on a

protein solution contain 1.15mg/L in 10mM phosphate buffer of pH 7, with 0mM

(—s—), 50mM (– –œ – –), or 150mM (- - -i- - -) sodium chloride. The rate

(in mg/m2) was calculated from the area increase by using the �–area isotherm of

spread monolayers of b-lactoglobulin. Data from Ref. 195, where the experimental

details are given.


the rate of incorporation of b-lactoglobulin into monolayers of DSPA,distearoylphosphatidylcholine (DSPC) and DPPA is shown versus surfacepressure (�) at pH 7. The rate was calculated using a simple first-orderkinetics model (8), where only the surface pressure barrier is taken intoaccount. The highest rate of adsorption of b-lactoglobulin into a phospho-lipid monolayer was observed for anionic DSPA. The incorporation of theprotein take also place at a higher surface pressure into a DSPA monolayerthan into a monolayers of the other lipids. Because the b-lactoglobulin, witha zero net charge at pH�5.2 (196), has a positive net charge at pH 4, a largerrate of adsorption into the negatively charged phosphatidic acid monolayerswould be expected under acidic conditions. However, almost the same rateswere found (195). As discussed earlier, anionic lipids seems to interact morestrongly with proteins (i.e., to its cationic amino acid residues) compared tolipids with no or positive net charge. The incorporation into the zwitterionicDSPC monolayers is as expected less salt dependent than what was observedfor the phosphatidic acid monolayers, where the rate increases with increas-ing ionic strength of the subphase. Probably, this is a consequence of adecreased repulsion within the phosphatidic acid protein monolayer at ahigher ionic strength. The findings by Bos and Nylander (195) is somewhatcontradictory to the findings of Cornell and Patterson, who studied theadsorption b-lactoglobulin to a negatively charged lipid monolayer,composed of a mixture of palmitoyloleoylphosphatidylcholine (POPC)and palmitoyloleoylphosphatidylglycerol (POPG) (65/35 mol%). Theyonly observed a substantial binding of b-lactoglobulin at pH 4.4, which iswhen the protein carries a net positive charge, but not at higher pH (pH 7)(197). The differences probably arises from the different lipids and metho-dology used by Cornell et al. (197–199). Cornell et al. measured the amountsof protein adsorbed to the lipid layer by transferring the layer to a solidsupport. During the transfer, the surface pressure was kept at 30–35mN/m,thus preventing insertion of portions of the protein in the lipid monolayer(199). Only protein molecules that interact strongly with the lipid headgroups are transferred to the solid supported. Another difference is thattheir surface pressure data of the protein penetration is recorded underconstant area, not at constant pressure, as in our study. In addition,Cornell et al. used lipids with their chains in the liquid state, which, aswill be discussed, can influence the interaction. Cornell (198) also observeda specific interaction between b-lactoglobulin and egg yolk phosphatidicacid (e-PA) in spread mixed films at low pH (1.3 and 4) where b-lactoglo-bulin carries a positive net charge. No interaction was observed for e-PAin the neutral pH range or for egg yolk phosphatidylcholine (e-PC). Similarobservations was made for the interaction between a-lactalbumin or BSAwith mixed monolayers of POPC and POPG, where adsorption was


observed below the isoelectric point of the protein, where the lipid layerand the protein carry opposite net charge, but less was adsorbed aroundand almost nothing above the isoelectric point (199). The inter-action was reduced in the presence of calcium as well as at increasedionic strength. Cornell et al. thus concluded that the interaction is of anelectrostatic origin.

The work of Quinn and Dawson concerning the interaction betweencytochrome-c (positive net charge below pH 10) and phospholipids from eggyolk also stresses the importance of the electrostatic interaction, althoughconformational changes of the protein are of importance (200,201). Theymeasured the pressure increase caused by the penetration/adsorption of theprotein to the lipid monolayers as well as the amount adsorbed by using 14C-labeled protein. Their results show that the limiting pressure for penetrationis 20 and 24mN/m for phosphatidylcholine and phosphatidylethanolamine,respectively, whereas penetration into the phosphatidic acid and diphospha-tidylglycerol (cardiolipin) monolayers occurred up to pressures close to thecollapse pressure of the film (<40mN/m). Furthermore, the penetrationinto the e-PC monolayers was not affected by increasing the sodium chlorideconcentration to 1M. Cytochrome-c bound to the e-PC monolayers couldnot be removed by increasing the ionic strength. This is in contrast to thecardiolipin and e-PA monolayers, where the penetration was reduced whenthe sodium chloride content was increased to 1M. It was also possible topartly desorb some cytochrome-c from e-PA monolayers. However, thepH dependence of the interaction was found to be quite complex, whichsuggests that subtle changes in the protein conformation also affect theinteraction.

The importance of the electrostatic interaction with the phospholipidhead group has also been shown by the work of Malmsten et al. (202,203),who studied the interaction of human serum albumin, IgG, and fibronectinfrom human plasma with phospholipid layers spin-coated onto methylatedsilica surfaces. Generally, he found no interaction between the proteins andlipids with no net charge or with shielded charges (e.g., phosphatidylcholine,phosphatidylethanolamine, sphingomyelin, and phosphatidylinositol),whereas interaction was observed with the surfaces containing unprotectedcharges (e.g., phosphatidic acid, diphosphatidylglycerol, and phosphatidyl-serine).

Hydrophobic Interactions. As observed in Fig. 8, the rate of adsorp-tion of b-lactoglobulin into DPPA monolayers was significantly lower thaninto the monolayers, where the corresponding lipid had a longer chainlength. This points to the importance of hydrophobic interactions for theincorporation. It was also observed that the incorporation was much faster


into the lipid monolayer than into its own proteinous layer, being less ‘‘oil-like’’ than the lipid layer (195). In addition, repulsive steric and electrostaticforces might contribute the lower rate of incorporation. Quinn and Dawson(201) found that the threshold surface pressure, above which no penetrationof cytochrome-c took place in phosphatidylcholine monolayers, was con-siderably lower when DPPA was used instead of hydrogenated e-PC. Thelatter lipid contained fatty acid with a longer chain length, about 60% C18

and 30% C16. Du et al. (204) studied the influence of the alkyl chain lengthof glucolipids (dialkyl glycerylether-b-D-glucosides and dialkyl glyceryl-ether-b-D-maltosides) on the interaction between lipid monolayers and glu-cose oxidase. The interaction, as shown by an increase in surface pressure,was found to increase with increasing lipid chain lengths for both types oflipid. These results suggest that the hydrophobic interaction is the predo-minant force. Furthermore, it is interesting to note that the interactions werenot so strong with the lipids having the more bulky head group (i.e., thedialkyl glycerylether-b-D-maltosides), although the �–A isotherms for thecorresponding dialkyl glycerylether-b-D-glucosides were similar. This illus-trates that a bulky head group can sterically hamper the protein–lipid(hydrophobic) interaction.

Effect of Lipid Fluidity. The complete hydrogenation of e-PC wasfound not to affect the surface pressure threshold for penetration of cyto-chrome-c compared to the native e-PC (201). However, the change in sur-face pressure due to the penetration of the protein versus initial surfacepressure was less steep for the saturated one. A similar trend was observedfor the e-PE samples (200). The conclusion was that the limiting pressure forpenetration to take place is likely to be determined by the work necessary forthe penetration, that is

R� dA, where an area, A, of the interface has to be

created for the protein to penetrate. Once the penetration is feasible, themagnitude will depend on the space between the molecules, and, thus, thedegree of penetration is expected to be lower for the hydrogenated sample(201). The surface pressure threshold below which penetration of cyto-chrome-c into the anionic diphosphatidylglycerol (cardiolipin) monolayertook place was also found to decrease when the lipid was fully hydrogenated(201). Ibdah and Phillips found the same trend in their study of the effect oflipid composition and packing on the penetration of apolipoprotein A-I intolipid monolayers (205). In the biological system, this protein interactswith the phospholipid membrane of the serum high-density lipoprotein(HDL) particles (see Section V.B). Their results show that this proteinadsorption occurs, to a larger extent, on expanded monolayers than oncondensed monolayers (i.e., protein adsorption decrease in the ordere-PC>egg sphingomyelin>DSPC). Furthermore, it was found that


protein adsorption generally decreased with an increasing amount of cho-lesterol in the lipid monolayer. It was suggested this was due to the conden-sing effect of cholesterol.

Effect of Protein Structure. Evidence of the importance of theprotein structure was provided by Hanssens and Van Cauwelaert (206),who studied the penetration of a-lactalbumin in monolayers of DPPCand cardiolipin at physiological pH (pH 7.4) and at pH 4.6 withand without calcium. Reports have indicated that the protein is trans-ferred to an intermediate, partially unfolded (molten globule) state atlow pH and by depletion of calcium (101,103,104), which would expectto facilitate the penetration of a-lactalbumin into the monolayer.Furthermore, it has been reported that the calcium-free form of a-lactal-bumin is more hydrophobic (207). Indeed, penetration occurred atlow pH and was prevented if the protein was adsorbed from a calciumsolution (206).

It is not easy to in situ monitor the structural changes of the protein onbinding to a lipid monolayer. However, they can be monitored indirectly byrecording the circular dichroism (CD) spectra of the mixed protein–lipidfilm transferred to a quartz plate. This technique was used to recordedCD spectra for b-lactoglobulin, a-lactalbumin, or BSA bound to mixedmonolayers of POPC and POPG (197,199). The spectra show that theprotein bound to the lipid monolayer was similar to the one recorded insolution, indicating that the conformation of the protein did not changesignificantly when interacting with the lipid monolayer.

2. Structure of the Interfacial Film

Even from the study of the penetration of protein versus surface pressure, itis also possible get some hints about the structure of the mixed layer. Cornellet al. (197,199) observed penetration of b-lactoglobulin, a-lactalbumin, orBSA into mixed monolayers of POPC and POPG at such high surfacepressure that it is unlikely that the proteins could penetrate into a proteinlayer. Thus, they concluded that the formation of pure protein patchesis unlikely and that portions of the protein is suggested to be inter-calated into the lipid monolayer. Bos and Nylander made similar observa-tion for the interaction between b-lactoglobulin and DSPC and DSPAmonolayers (195).

Fluorescence microscopy and Brewster-angle microscopy (BAM) canbe used to in situ image the structure of the film at the air–aqueous interface,although the lateral resolution is limited by the resolution of the opticalmicroscope. Fluorescence microscopy together with the surface filmbalance technique was used by Heckl et al. to study the structure of


mixed phospholipid–cytochrome-c and cytochrome-b films (208). Theyfound that proteins primarily were located in the fluid membrane phase,which coexisted with solid lipid domains without protein. The penetrationinto the lipid monolayer was reduced with increasing pressure. Cytochrome-c (positively charged) was found to interact with dimyristoylphosphatidicacid (DMPA) monolayers but not with dipalmitoylphosphatidylcholine(DPPC) layers, showing the electrostatic nature of the interaction.Schonhoff et al. concluded from their study of the incorporation of mem-brane proteins into DPPA/DOPA monolayers that incorporation mainlytakes place in the fluid phases of the matrix (209). Zhao et al. used BAMto image the kinetics of b-lactoglobulin penetration into DPPC monolayersat the air–aqueous interface from a 500-nM solution in 10mM phosphatebuffer, pH 7 (210). For instance, at an initial surface pressure of 7.8mN/m,it took 0.17min until domains, with similar morphology as those appearingduring the compression of a pure DPPC monolayer, appeared. Thesedomains were found to consist only of the lipid, as confirmed by grazingincidence x-ray diffraction, and b-lactoglobulin penetration was found tooccur without any specific interaction with DPPC. b-Lactoglobulin was notable to penetrate into a condensed DPPC monolayer (i.e., above a surfacepressure of about 20mN/m).

The lateral organization in mixed protein–lipid films at the air–aqueous interface can be studied by spectroscopic techniques and high-resolution imaging techniques such as electron microscopy and atomicforce microscopy (AFM), after transferring the films to a solid support.Using electron microscopy, Cornell and Carroll found that only lipidswith the chains in liquid state (e-PA, dioleoylphosphatidylcholine, anddioleoylphosphatidylethanolamine) formed homogenous films with b-lac-toglobulin, whereas DPPA and DSPC formed heterogeneous layers (211).The use of AFM as powerful technique for studying the lateral organiza-tion in mixed films of proteins and soluble surfactants has already beendemonstrated with the development of the ‘‘orogenic’’ displacement model(173–175). Diederich et al. studied the interaction between bacterial sur-face layer proteins (S-layer proteins) and phosphatidylethanolamine(DMPE and DPPE) monolayers using dual-label fluorescence microscopy,Fourier transform infrared (FTIR) spectroscopy, and electronmicroscopy(212). When the monolayer is in the two-phase region, with one isotropicand one anisotropic fluid phase, the S-layer protein adsorbed preferen-tially to the isotropic phase. However, two-dimensional crystallizationcould be nucleated in the boundaries between the two phases, but pro-ceeded mainly underneath the anisotropic phase. The FTIR measurementsclearly indicate that the protein crystallization leads to an increased orderof the lipid acyl chains.


3. Monolayer Stability

One might expect that monolayer made up of lipids with a very low aqueoussolubility would be stable. However, this is far from general. The metasta-blility of monolayers can be caused by processes such as rearrangementwithin the layer, dissolution into the subphase, and transformation to athree-dimensional phase, which can occur at pressures above the equili-brium spreading pressure (213,214). Furthermore, the stability of the mono-layers can be affected by the spreading solvent and the techniques used forspreading the lipid (215,216). The stability of the monolayer can also beconsiderably changed by the ion composition of the aqueous subphase.For instance, the stability of an arachidic (n-eicosanoic, C20:0) acid mono-layer was found to increase in the order Hþ<Liþ<Naþ<Ca2þ<Mg2þ

(213). There are several examples of proteins that are thought to have therole to stabilize a lipid monolayer or bilayer. One such example is the milkfat globule membrane that has been suggested to consist of the monolayer ofpolar lipids, which covers the fat globule surface, and an outer lipid-basedbilayer (217,218). The milk fat globule membrane is expected to be inho-mogeneous, with a significant amount of proteins in the membrane. Anaqueous layer containing different proteins, like xanthine oxidase, is presentbetween the monolayer and bilayer. One of the roles that have been assignedto xantinoxidase is to stabilize the milk fat globule membrane (218).Interestingly, Kristensen et al. found that the presence of a xanthine oxidasecan increase the stability of a monolayer composed of sphingomyelin fromthe milk fat globule membrane (219). They investigated the interactionbetween one of the major proteins, xanthine oxidase, and the majorlipids, sphingomyelin and phosphatidylcholine, in the milk fat globule mem-brane at the air–aqueous interface by using the monolayer technique. Bothlipids have a similar phopshorylcholine head group, which is zwitterionic inthe neutral pH range, although the belt regions linking the phopshoryl-choline group with the acyl chains are different. The �–A isotherms ofsphingomyelin and phosphatidylcholine are shown in Figs. 10a and 10b,respectively. The isotherms for sphingomyelin monolayers spread on purebuffer and a xanthine oxidase solution are shown. The slope of isotherm andthe area of the compressed monolayer for pure sphingomyelin (Fig. 10a) aresmaller than expected for these types of lipid. In addition, the large hyster-esis and the dependence on the compression speed, not observed for dis-tearoylphosphatidylcholine, confirms that the sphingomyelin monolayer ismetastable. The difference in stability of monolayers formed by two differ-ent lipids can probably be related to the different conformations of cholinegroups in the two types of lipid, where intramolecular hydrogen-bonding ispossible between the phosphate group and the amide and hydroxyl groups


in the belt region of sphingomyelin (220). An increase in � at maximumcompression of the sphingomyelin monolayer, which reflects an increase inthe monolayer stability, was observed in the presence of sphingomyelin.Furthermore, the area per sphingomyelin molecule increases in the presenceof xanthine oxidase even at high � values. This is in contrast to the resultsfrom the parallel study of the phosphatidylcholine monolayers with andwithout xanthin oxidase, where the interacting protein could be completely

Figure 10 Dynamic surface pressure (�) as a function of the molecular area of the

spread amount lipid for compression of (a) sphingomyelin and (b) DSPC monolayers

on a phosphate-buffered subphase (40mM phosphate containing 0.1M sodium

chloride, pH¼ 7.4) with or without xanthine oxidase (5mg/mL). The isotherms

recorded for the lipid spread on pure buffer (—) and at 5 (- - -), 10 (– – –), and 20

(– - –) min elapsed between spreading and compression. The lipid (25 mg) was spreadfrom a chloroform/methanol (2:1, v/v) solution on a maximum area of 50� 450mm2

and a compression speed of 12.5mm/min was used. (Data from Ref. 219, where the

experimental details are given.)


squeezed out from the lipid monolayer at high enough surface pressureswithout affecting the collapse pressure. This indicates that the interactionbetween xanthine oxidase and sphingomyelin is much stronger than thatbetween the protein and phosphatidylcholine.

B. Protein-Lipid Interactions at the Oil–Aqueous Interface

Most studies of protein–lipid interactions at the oil–aqueous interface hasbeen carried out using model emulsions. The purity of polar lipid and theway it is added (e.g., to the oil or the water phase) are bound to affect theinteractions with proteins, which, in turn, affect the emulsion stability.Yamamoto and Araki (221) studied this by comparing the interfacial beha-vior of b-lactoglobulin, in the presence of lecithin (PC) in the water or in theoil phase, with the stability of corresponding emulsions. In the presence ofprotein, crude lecithin was found to increase the stability of emulsion andlower the interfacial tension more effectively than a pure lecithin prepara-tion. When crude lecithin was added to the oil phase, the interfacial tensionwas found to decrease and the emulsion stability increased compared towhen the lecithin was dispersed in the aqueous phase. One might speculatewhether these findings can be related to the presence of fatty acid and/orcharged phospholipids in the crude lecithin. Aynie et al. studied the inter-action between nitroxide homologs of fatty acids and milk proteins byfollowing the mobility of the nitroxide radicals using electron spin resonance(222). At pH 7, the importance of the lipid–protein interaction was notdetermined by the structure of the protein, but positively correlated withthe number of positive charges on the protein. Thus, it was observed that theimportance of the interaction in the emulsions decreased in the orderas1-casein> b-lactoglobulin> b-casein, suggesting that the interaction wasof an electrostatic nature. The different proteins also affect the organizationof the lipid monolayer, where as1-casein, in contrast to b-lactoglobulin andb-casein, induce, an ordering of a monolayer of nitroxide fatty acids on thesurface of an emulsion droplet (222). This can probably be assigned to thestronger interaction of as1-casein with lipids compared to the other proteins.

Bylaite et al. studied the stability and droplet size of b-lactoglobulin-and lecithin [phosphatidylcholine from soybean (sb-PC)]-stabilizedemulsions of caraway essential oil as well as the amount of protein on theemulsion droplets (223). It should be noted that sb-PC was dispersed in theoil phase. Some of their data are given in Fig. 11, where the amount ofb-lactoglobulin adsorbed on the oil–aqueous interface is shown versusamount added of s-PC. These data show that sb-PC is likely to replacesome of the protein at the oil–aqueous interface, although it is unable tocompletely replace the protein. The maximum reduction in the amount of


b-lactoglobulin adsorbed is by a factor of 3 for the caraway oil. Thesefindings are in agreement with other studies, where lecithin was found tobe less efficient in displacing milk proteins from the oil–water interfacecompared to other surfactants (224,225). Bylaite et al. found that emulsionswith triglyceride oil generally proved to be more stable compared to thosemade with caraway essential oil as the dispersed phase (223). However, thestability of the emulsions could be improved considerably by adding sb-PC.An increase in the protein concentration also promoted emulsion stability.Fang and Dalgeish arrived at a somewhat different conclusion for casein-stabilized emulsions (226). They found that the presence of DOPC destabi-lized casein-stabilized emulsions of soybean oil in 20mM imidazole/HCl atpH 7.0. This seemed to be independent on whether DOPC was presentduring emulsification or if it was added to the emulsion as dispersed aggre-gates. At a high concentration of casein, the emulsions were stable, and thedecrease in surface load was a direct indication of the removal of caseinfrom the interface by the presence of DOPC. The higher the DOPC con-centration, the greater was the effect on emulsion stability and surface load.DPPC and egg PC either enhanced or did not affect the stability of theemulsion.

Bylaite et al. applied ellipsometry to study the adsorption of the lipidfrom the oil and the protein from the aqueous phase at the oil–water interface(223). Independently of the used concentration, close to monolayer coverage

Figure 11 Adsorbed amount of protein at the caraway essential oil–water (i, �)

and olive oil–water (s,œ) interfaces in emulsions stabilized by 1 (i,s) and 2 (�œ)

wt% b-lactoglobulin and variable amount of soybean-PC. Emulsions were prepared

from 15wt% oil in a 60mM phosphate buffer of pH 6.7. (Data Ref. 223 where the

experimental details are given.)


of sb-PC was observed at the caraway oil–aqueous interface. On the otherhand, at the olive oil–aqueous interface, the presence of only a small amountof sb-PC led to an exponential increase of the layer thickness with timebeyond monolayer coverage. This interesting observation was assigned tothe formation of a multilamellar layer of sb-PC at the olive oil–aqueousinterface, when sb-PC reached the solubility limit in the olive oil.

The displacement of caseinate from the interface of emulsion dro-plets by monoglycerides, mono-oleoylglycerol, and monostearoylglyceroldissolved in the oil phase was found to correlate with the adsorption ofthe monoglycerides at the oil–water interface (227). The amount of mono-oleoylglycerol increased gradually with concentration and reached a pla-teau when approaching an oil-phase concentration of 1wt%. Under theseconditions, all of the caseinate was displaced from the interface. Thesaturated lipid, monostearoylglycerol, was much more efficient in displa-cing the protein. Already, at a concentration in the oil phase of between0.2 and 0.3 wt%, the adsorbed amount of monostearoylglycerol increasedsharply and reached much higher surface concentrations than mono-oleoylglycerol. At 0.3 wt%, all of the caseinate was removed from theinterface.

C. Protein Interactions with Lipid Vesicles

The mechanisms that determine the stability, size, and shape of vesicles arecomplex and widely discussed (for reviews, see, e.g., Refs. 80–83). Thespherical shape is generally the most stable shape for equal distributionof molecules between the two monolayers constituting the bilayer (81).These spherical vesicles can be large multilamellar vesicles (MLV) andlarge (LUV) and small (SUV) unilamellar vesicles (81). The bending ofthe lipid bilayer to form a vesicle imposes a strain on a symmetric bilayer,as the inner monolayer has a negative curvature, whereas the outer has apositive curvature. The magnitude of this curvature energy can be difficultto estimate, but it is thought to be significant enough in many cases tomake the vesicles inherently unstable and energy has to be added to formthem (81–83). The result of the tension can be a nonspherical vesicle (228).A mixture of phospholipids, which corresponds to the composition in themilk fat globule membrane, gives both spherical vesicles and tubular struc-tures (229). In particular compositions (e.g., 80% DOPE, 12% DOPC, and8% sphingomyelin) that at high lipid content give liquid-crystalline phasesat the boundary of lamellar to reversed hexagonal phase tend to givemicrotubular structures at high water content rather than vesicles. Alarger proportion of multilamellar vesicles were observed in buffer anddivalent salts than in pure water. A small increase in the interlayer spacing


of the mulitlamellar vesicle was observed in the presence of b-lactoglobulinand b-casein.

1. Driving Force for the Protein–Vesicle Interaction

The driving mechanism for the interaction of proteins with the lipid bilayerof the vesicles are basically as for the interaction a lipid monolayer at theair–aqueous interface. In parallel to the Quinn and Dawson study discussedearlier (200,201), Rytomaa et al. (230) found a strong electrostatic contribu-tion when cytochrome-c binds to cardiolipin–phosphatidylcholine lipo-somes. This interaction did not take place if the negatively charge lipidcardiolipin was absent in the membrane. Furthermore, the protein was dis-sociated from the vesicle in the presence of 2mM MgCl2 and 80mM NaClat pH 7. The apparent affinity of cytochrome-c to the vesicles also increasedwhen the pH was dropped to 4. The interaction was found to be completelyreversible for pH changes; that is, if the pH was increased to 7, the proteincould be dissociated from the vesicle by adding salt.

Price et al. studied the adsorption of fibrinogen to neutral liposomes,composed mainly of phosphatidylcholine (PC) and cholesterol and negativeliposomes, composed mainly of phosphatidic acid (PA) and cholesterol, aswell as to the corresponding liposomes in which a polyethylene glycol(PEG)-modified phosphatidylethanolamine had been introduced (231).They found that negatively charge liposomes adsorbed more fibrinogenthan the corresponding neutral liposomes. PEG modification was foundto have no effect on neutral liposomes in terms of fibrinogen adsorption.However, PEG modification, which sterically stabilizes the liposome, mark-edly reduced the adsorption to the negative liposomes.

Brooksbank et al. conducted an extensive study on the interaction ofb-casein, k-casein, as1-casein, and b-lactoglobulin with negatively chargede-PG and zwitterionic e-PC vesicles using photon correlation spectroscopy(232). Their data on the adsorption of b-casein are shown in Fig. 12. All ofthe studied proteins were found to give a thicker layer on the negativelycharged vesicles, although they all carried a negative net charge under theconditions used (160mM sodium chloride at pH 6.2). Brooksbank et al.(232) suggested that binding to the vesicle surface takes place mainlythrough hydrophobic interactions and the differences in thickness of theadsorbed layers on the two types of vesicles were explained in terms ofthe protein charge distribution. For instance, the hydrophilic, N-terminal,part of b-casein has a net charge of �12, whereas the remainder of themolecule carries almost no net charge. Thus, on the negatively charge vesiclesurface, the molecules adopt a more extended configuration, as the N-ter-minal part is likely to be pushed away from the surface by means of


electrostatic repulsion. This explains the thicker layers on this surface asshown in Fig. 12. A similar reasoning can be applied for k-casein. Theapparently very thick adsorbed layer of as1-casein was explained by bridgingflocculation of the vesicles mediated by the protein. The middle section ofas1-casein carries a negative net charge, whereas the two ends have no netcharge. One of the uncharged ends pertrudes into the vesicle bilayer and themiddle section is repelled from the vesicle surface, leaving the otheruncharged end of the peptide chain free to interact with another vesicle.The charge distribution on b-lactoglobulin is more even and the interpreta-tion of the results was not as straightforward.

As discussed by Kinnunen, the introduction of the HII-formingdouble-chain lipid (a lipid with packing parameter >1; see Fig. 1) in alamellar membrane imposes a considerable stress on the membrane (1).This frustrated membrane is said to be in the Le state according to theKinnunen terminology (1). Free energy can be gained by allowing some ofthe lipids in the frustrated membrane to adopt the so-called extended orsplayed chain conformation, where one of the acyl chains extends outfrom the bilayer, whereas the other chain remains in the membrane.Such an extend chain can also be accommodated within a proper (hydro-phobic) cavity of protein interacting with the membrane (1). This is aninteresting alternative explanation for the hydrophobic interactionbetween peripheral proteins and membranes that has been discussed inthis review. The splayed chain conformation has also been suggest to beone mechanism for membrane fusion (2). This and other implications of

Figure 12 Thickness of adsorbed layer on b-casein on negatively charged e-PG

and zwitterionic e-PC vesicles as a function added protein expressed as microgram of

protein per square meter of available liposome surface. The liposomes were dispersed

in 160mM and the pH was about 6.2. The data are taken from a photon correlation

spectroscopy study by Brooksbank et al, where further experimental details are

given (232).


the splayed chain confirmation has recently been discussed by Corkery(233).

2. Influence of the Protein Structure on the Vesicle Interaction

Kim and Kim studied the interaction between a-lactalbumin and phospha-tidylserine–phosphatidylethanolamine vesicles (1:1 molar ratio) versus pH(234). They found that the interaction, which almost did not exist at neutralpH, increased with decreasing pH (Fig. 13). What is interesting to note(Fig. 13) is that vesicle fusion, as estimated from the increase of the initialrate of Tb fluorescence increase, correlates with the binding of the protein tothe vesicles. The binding was suggested to be due to hydrophobic interactionvia protein segments penetrating into the lipid bilayer, as it was impossible todissociate it by increasing the pH. This was further confirmed by usingproteolytic enzymes, which were found to cut off both ends of the polypep-tide chain, leaving only the segment that penetrates into the bilayer. Thispenetration protein loop was also believed to induce fusion of the vesicles.

Figure 13 The initial rate of Tb fluorescence increase (- - -s - - -, - - -œ - - -) upon

a-lactalbumin induced fusion of phophatidylserine–phosphatidylcholine (1:1 molar

ratio) vesicles is shown as a function of pH. The pH-dependent binding of

a-lactalbumin is shown as the amount of protein bound per milliliter of vesicle

suspension (—f—, —g—), which contained 1mM lipid molecules (determined

from the phosphor content) per milliliter of suspension. The results for initial protein

concentrations of 50 (s, f) and 100 (œ, g) mg/mL are presented. As the curves for

the fusion process represents kinetic data and the binding studies represent

equilibrium data when the fusion process is over, only a qualitative comparison is

possible. (Data Ref. 234, where the experimental details are given.)


The importance of the protein conformation on the interaction withvesicles was also shown in the study of Brown et al. (235). They found nointeraction between native b-lactoglobulin and DPPC vesicles, but b-lacto-globulin, modified by exposing it to a 2:1 mixture of chloroform and metha-nol, did interact with the vesicles. Moreover, the lipid–protein complexformed had an a-helix content of at least 25–30% larger than for thenative protein. The interaction was found to lead to aggregation of thevesicles at pH 7.2, whereas no aggregates were observed at pH 3.7. Thiswas explained by the larger net charge at pH 3.7 (þ20) compared to pH7.2 (�10). These results imply that protein modification, either during pro-cessing or by special treatment, can increase the helix content, which, in, turncan be boosted by lipid interaction. The lipid–protein complexes formed havebeen suggested as a way to improve the emulsification processes (236,237).

3. Lateral Phase Separation in Vesicle Bilayers

Raudino and Castelli reported that the presence of lysozyme could inducedlateral phase separation in vesicle bilayers composed of a mixture of phos-phatidic acid and phosphatidylcholine (238). Their differential scanningcalorimetry study of the lipid chain melting transition showed goodmixing in the absence of the protein and the single peak was shiftedtoward higher temperatures as the phosphatidic acid content increased. Inthe presence of lysozyme however, lateral phase separation did occur as thechain melting transition peak was split into two peaks. In addition, theyfound that the temperature of protein unfolding increased with the fractionof phosphatidic acid, suggesting a stabilization of the protein due to theinteraction with phosphatidic acid.

It is important to bear in mind that microheterogeneity of the bilayernot only occurs for mixtures of different lipids but also close to the gel-to-fluid phase transition of the lipid. Hønger et al. studied the relation betweenphospholipase A2-catalyzed hydrolysis of one-component phosphatidylcho-line vesicles and the microheterogeneity of the lipid bilayer (239). Theyvaried the microheterogeneity by changing the temperature in the vicinityof the gel-to-fluid phase transition as well as using lipid chain lengthsbetween C14 and C18 and found a strong correlation between the maximallipase–lipid interaction and the maxima in the interfacial area between geland fluid domains.

D. Protein Interaction with Liquid-Crystalline Phases/Gel Phases

As discussed previously, the interaction between water-soluble lipids andproteins usually takes place via monomers, and the associated surfactant


structures is generally formed in the same range, or higher, of surfactantmonomer concentration as needed for the interaction to take place. Forlipids with low aqueous solubility, on the other hand, the association struc-tures, different lyotropic mesophases, are generally already present whenmixed with the protein and have thus a profound impact on the protein–lipid interaction. It is therefore important to be familiar with the phasebehavior of the particular lipid used because often seemingly conflictingresults can be derived from differences in the phase structure of the lipids.For instance, we have observed that the interaction between b-lactoglobulinand phosphatidic acid only occurred when the lipids were present asa dispersion, but not when they were mixed in the gel state with theprotein (22).

Even if no specific interaction occurs, proteins can have an impact onthe liquid-crystalline phase or gel phase due to the limited space of theaqueous cavity. This was demonstrated by Minami et al., who investigatedthe incorporation of lysozyme, b-lactoglobulin, and a-lactalbumin in asphingomyelin gel phase containing 0.6wt% sodium palmitate and80 wt% aqueous solution (240). The dimension of the aqueous layer inthe gel phase was suggested to limit the amount of protein that could beincorporated. Above this limit, phase separation will occur with a gel phaseand an ‘‘outside’’ protein-rich solution. The protein will, at high enoughconcentration, probably also compete for the water in the interlamellarspacing, which will eventually lead to a reduction of the aqueous layerthickness. This effect was demonstrated for high-molecular-weight polymersin equilibrium with the phosphatidylcholine lamellar phase (241). The poly-mer was unable to enter the aqueous layer, but still exerted an osmotic stressthat was large enough to compress the lamellar lattice, as shown by x-raydiffraction data. This method has been used to measure the interactionbetween the lipid bilayers (241,242).

Proteins are, of course, also able to enter into the aqueous layer of alamellar phase and thereby affect the swelling. This was shown by Rand(243), who studied the penetration of bovine serum between negativelycharged lecithin/cardiolipin bilayers at pH 3.3, where the protein has apositive netcharge. BSA is also likely to adopt a more expanded structureat this pH, thus exposing more hydrophobic segments. He found that theinterlamellar spacing of the lamellar phase, decreased with an increasingcardiolipin/bovine serum albumin ratio. This was related to a reduction ofthe negative charge of the lipid layer as the amount of bound proteinincreases.

We will start our discussion by giving some example of the interplaybetween the lipid structures and protein in terms of the effect on the curva-ture of the lipid–aqueous interface, because curvature has an important role


in condensed matter, as discussed in the book by Hyde et al. (59). McCallumand Epand found that changing the curvature of biological membranescould modify membrane-bound insulin receptor autophosphorylation andsignaling (244). This was demonstrated by adding compounds that raisedthe bilayer to the reverse hexagonal (HII) transition temperature of modelmembranes. This promoted a decrease of curvature and inhibited the insulinstimulation of the receptor phosphorylation.

The next part will focus on a specific structure, the bicontinuous cubicliquid-crystalline phase, where the interaction with the lipid bilayer as wellas the physical entrapment within the cubic phase takes place. Finally, wewill briefly discuss the effect of enzyme interactions, namely lipase action, onliquid crystalline phases.

1. Protein Interactions That Increase the Curvature ofthe Lipid–Aqueous Interfaces

Proteins or peptides that penetrate into the hydrophobic domain of the lipidbilayer generally provoke an increase of curvature of the lipid–aqueousinterface (i.e., becomes more concave toward the aqueous space). Quite afew of the membrane-bound peptides have these properties, such asGramicidin A, a hydrophobic polypeptide, which forms channels for mono-valent cations in phospholipid membranes (245). This peptide was found tofavor the transition lamellar phase !reversed hexagonal (HII) phase indioleoylphosphatidylcholine (DOPC) and dioleoylphosphatidylethanola-mine (DOPE) systems in an excess of water, as observed by nuclearmagnetic resorance (NMR) studies (246).

Not only can proteins or peptides that penetrate into the lipid bilayerinduce phase transitions, but also proteins that are certain to interact withthe head groups of the phospholipid bilayer can give rise to similar effects.This has been demonstrated for cytochrome-c, which has a positivenetcharge and has been shown to interact with negatively chargedphospholipids (247). The binding of cytochrome-c to anionic cardiolipinliposomes induced the formation to an inverted hexagonal, HII, structure(247). No interaction and, hence, no phase transition was observed in thepresence of liposomes composed of neutral zwitterionic lipids like PC andphosphatidyl ethanolamine (PE). A phase transition to the HII-phase wasobserved if a sufficient fraction of these lipids was replaced for cardiolipin,Interestingly, the protein was found to interact with liposomes with theanionic lipid phosphatidylserine (PS), but did not induce any phase transi-tion. The interaction between cardiolipin and cytochrome-c was also stu-died by Spooner and Watts, using deuterium and phosphorus-31 NMRmeasurements (248). Likewise, they found that the interaction can, depending


on the lipid stoichiometry, cause a transition from a lamellar to a non-bilayer structure. The binding of the protein with the liquid-crystallinebilayers of cardiolipin was also found to cause extensive derangement ofthe cytochrome-c secondary structure (248,249).

Studies of the interaction between cytochrome-c and suspensions ofDMPG or admixtures of dioleoylglycerol (DOG) or DOPC with DOPGalso showed that binding of cytochrome-c could promote an increase insurface curvature of the lipid aggregates from a bilayer structure (250).This is deduced from NMR data where an isotropic peak occurs in thepresence of cytochrome-c, indicating cubic lipid phases, small sphericalvesicles, or extended bilayers with high local curvature. The structure ofcytochrome-c was found to change on binding to the lipid, and twoforms, depending on the lipid composition, were identified with resonanceRaman measurements:

I. Close to the native conformation in solutionII. Unfolded with the heme crevice opened

The changes in protein structure could be correlated with the curvature ofthe lipid bilayer, as illustrated in Fig. 14, as the ratio between the unfolded(II) and native (I) cytochrome-c (cyt c) in DOPC/DOG dispersions versusDOG mol%. The presence of DOG was found to induce spontaneous cur-vature in the DOPG lipid bilayer in the pure lipid system, which at a DOGcontent of � 50%, leads to the transition to a reversed hexagonal (HII)

Figure 14 Concentration of unfolded (II) and native (I) cytochrome-c (cyt c) in

DOPC / DOG dispersions versus DOG mol% determined from Raman resonance

spectra. The concentrations of lipid and cytochrome-c were 300 and 20 mM,

respectively, in an aqueous buffer (1mM HEPES, 1mM EDTA) of pH 7.5. (Data

Ref. 250, where the experimental details are given.)


phase. In the absence of DOG (i.e., a strict bilayer structure), the binding ofthe more unfolded form (II) of cytochrome is favored, whereas the fractionof the more native globular protein structure (I) increases with the amountof DOG (Fig. 14) and thus with curvature of the surface. The physical stateof the lipid was also found to affect the proportions of the two structuralforms of cytochrome-c. In the fluid state of pure DMPG, the fraction of themore unfolded form (II) was larger (85%) than when the lipid was in the gelstate (80%). It is noteworthy that they found that the bound fraction of themore unfolded form (II) to the fluid DOPG bilayer structure was substan-tially lower (75%), indicating that not only the fluidity of the bilayer mattersbut also the type of lipid.

The interaction between cytochrome-c and mono-olein in the cubicphase was studied by Razumas et al. by differential scanning calorimetry(DSC) and optical microscopy (251). In line with the above-reported studies,they also found that the presence of cytochrome-c at high enough concen-trations favored lipid aggregates with a larger curvature. Thus, theyobserved that the phase transition cubic ! HII ! L2 in the mono-olein–cytochrome-c–water system took place at a lower temperature than in thebinary mono-olein–water system (251). Similar effects were observed whenglucose oxidase was included into mono-olein–aqueous cubic phase (252).The temperature of the phase transition cubic ! HII in the mono-olein–glucose oxidase aqueous system decreases with the increase in glucoseoxidase concentration.

2. Protein Interactions That Decrease the Curvature ofthe Lipid–Aqueous Interfaces

Fraser et al. investigated the ability of a range of basic proteins and poly-lysine to convert a reversed hexagonal (HII) phase, consisting of DOPE andmixtures of DOPE and PS, to stable lamellar (La) phases at pH 9, whereDOPE is anionic, and at pH 7, when it is zwitterionic (253). The proteinsinvestigated were all capable of binding to the HII phase at pH 9, but onlymyelin basic protein and polylysine did induce transition to the La phase.Lysozyme formed a new HII phase in which the protein was included. Alowering of the pH seemed to release the proteins, except for mellittin, whichalso seemed to penetrate into the hydrophobic core of the lipid aggregates.The presence of PS in the HII phase at pH 7 increased the protein binding,but only interaction with myelin basic protein gave a lamellar phase. Basedon earlier studies, Fraser et al. suggested that the myelin basic proteinstabilized the lamellar phase by interacting with the DOPE headgroupand thereby increasing its effective size (253). They concluded that theproperties of myelin basic protein in terms of stabilizing the lamellar


structure could be related to the role of the protein to stabilize the myelinsheath multilayers.

3. Cubic Lipid–Protein Aqueous Phases

Lipid bilayers can be folded into such intriguing structures as cubic liquid-crystalline lipid–aqueous phases. In these structures not only the surfaceproperties and the curvature but also the dimensions of the aqueous spaceaffect protein–lipid interactions. The main features of the bicontinuouscubic phase is illustrated in Fig. 15. The mono-olein–aqueous system is athoroughly studied example of such a system, where two types of cubicphase have been observed on the water-rich side of the lamellar phase(84,254–257). We have already discussed the biological implications of lipid-cubic phases. Here, we will highlight some of the main features that are ofimportance for the functionality of lipid cubic phases. First, it is the bicon-tinuity of the cubic phase. This is illustrated Figs. 16a and 16b, where themobility of glucose solubilized in the aqueous channels and vitamin Ksolubilized in the lipid bilayer, respectively is illustrated. Figure 16a showsthe concentration profiles of glucose in the cubic mono-olein–aqueous phaseequilibrated against water as determined by holographic laser interfero-metry (258). These profiles could be fitted to Ficks second law, whichgave a diffusion coefficient four times lower than the value in the aqueoussolution. The mobility of the molecules in the aqueous channels of the cubicphase is certain to be affected by the dimensions of the channels and the sizeof the solute. Thus, electrochemical studies of the transport of cytochrome-cin the mono-olein–aqueous cubic phase gave values of diffusion coefficientsthat were about 70 times lower than the bulk values (251). Figure 16b showsthe mobility of mono-olein and vitamin K1 dispersed in the lipid bilayer asthe NMR self-diffusion coefficients plotted versus lipid volume fractionin the cubic phases. It is noteworthy that the mobility of the introducedvitamin K1 follows that of mono-olein, indicating complete dispersion ofvitamin K1.

The dimensions of the water channels in the bicontinuous cubicphases, which depend on the degree of swelling and type of cubic phaseare in the same range as the size of proteins (cf. Ref. 252). Furthermore,as liquid-crystalline phases they are quite flexible structures. These fea-tures have triggered a number of studies, which have shown that a largerange of hydrophilic proteins with molecular weights up to 590 kDacan be entrapped in the aqueous cavity of the mono-olein–aqueouscubic phases (251,252,259–262). The entrapped proteins have been foundto be protected in the cubic phase, with retained native confirmation(260,262–265). Enzymes can be kept for a very long time (months, in


Figure 15 Schematic of the cubic phase, showing the two unconnected water channel systems. The location of

the lipid bilayer is also indicated. The type of cubic phase shown can be described as a diamond of an infinite

periodic minimal surface (IPMS) corresponding to a primitive cubic lattice (Pn3m).


Figure 16 (a) Glucose concentration profiles in a mono-olein–aqueous cubic phase

(62:38wt%), where the aqueous solution initially contained 3.5wt% glucose, after

3 h (f) and 4 h (s) equilibration against pure water. The concentration is given as

the wt% glucose in the aqueous solution of the cubic phase. The solid and broken

lines represent the best theoretical fit of Fick’s law, giving diffusion coefficients of

1.39� 10�10m2/s and 1.47� 10�10m2/s after 3 and 4 h, respectively. The correspond-

ing bulk value is 6.7� 10�10m2/s. The data, obtained by holographic laser

interferometry, are from Refs. 258 and 261, where the experimental details are

given. (b) NMR self-diffusion coefficients at 25�C in mono-olein–aqueous cubic

phases containing 0–5wt% vitamin K1 are shown as a function of the lipid volume

fraction (including vitamin K1). The self-diffusion coefficients were measured in the

cubic (both gyroid and diamond type) and in the reversed micelle, L2, phases. Self-

diffusion coefficients of mono-olein (DMO) (f) and vitamin K1 (DVK) (s) are

shown. The lines are arbitrary fits to demonstrate the similar trends. The data are

from Ref. 271, where the experimental details are given.


some cases), with retained activity, which is not possible in the aqueoussolution (259,261).

Spectroscopic data have revealed changes in the molecular organiza-tion of the lipids evoked by the presence of the protein. FTIR measurementson the mono-olein–cytochrome-c aqueous system showed that the presenceof cytochrome-c increased the conformational order of the mono-olein acylchain and caused structural rearrangements in the polar-head-group region(251). These observations are in agreement with the decrease of the mono-olein packing parameter upon incorporation of cytochrome-c, which wasdeduced from the increase in unit-cell dimension of the cubic phase asdetermined by small-angle x-ray diffraction. The Raman scattering studieson the mono-olein–lysozyme–aqueous system demonstrated an increase inthe number of hydrogen -bond C¼O groups of mono-olein, but no increasein the acyl chain order relative to the binary lipid–aqueous system (260). Asimilar increase in the hydrogen-bonding, caused by the presence of a pro-tein, was observed in the mono-olein–hemoglobin–aqueous system usingFTIR spectroscopy (262). However, in this case, the protein incorporationalso caused a decrease in the acylchain order.

The cubic monoglyceride phases have also the ability to solubilizelipophilic proteins like A-gliadin from wheat (266) and bacterio-rhodopsin (267), as well as relatively large amounts of membrane lipids(260,261,268–270) and other hydrophobic compounds of biological rele-vance (268,271,272). These compounds are most probably dispersed in thelipid bilayer region of the cubic phase.

Apart from the biological significance of cubic lipid–aqueous phases,Razumas et al. demonstrated that cubic mono-olein–aqueous phases, con-taining enzymes, could be used as the biocatalytic layer in amperometricand potentiometric biosensors (259). Their results for biosensors, based ona variety of enzymes, show that the long-term stability decreases in theorder lactate oxidase>creatinine deiminase>glucose oxidase>urease,(i.e.; is basically in the order of increasing molecular weight). Also, thecubic phases of other amphiphiles like ethoxylated fatty alcohols can beused to entrap glucose oxidase, to construct a simple glucose monitor (60).The bicontinuous cubic structures have, by virtue of their well-definedporosity, also a large potential in drug delivery systems (46). Stable par-ticles of lipid–aqueous cubic phases, cubosomes, can also be produced forthis purpose (46,56,75,84–86). Landau and Rosenbusch demonstrated thatthe bicontinuous phases based on mono-olein and monopalmitolein couldprovide matrices for the crystallization of membrane proteins like bacterio-rhodopsin (267). They pointed out that the use of these types of cubicphase is advantageous as they provide nucleation sites, and the membraneproteins can be dissolved in the lipid bilayer. In addition, they support


growth by allowing lateral diffusion of the protein molecules in themembrane.

4. Lipase Action on Liquid-Crystalline Phases

Another food application, in which cubic lipid phases are believed to play animportant role, is in the degradation of lipids. This is also an example ofprotein–lipid interaction. However, lipases act at such a low concentrationthat their presence as protein does not significantly affect the global lipidself-assembly structure. It is rather their catalytic activity that has an impacton the lipid self-assembly structure. It is also important to remember thatthe action of lipases only decreases the time taken to reach the equilibriumand does not affect the equilibrium composition as such. Thus, the changesin structure and composition would have occurred even without the lipase ifgiven enough time. In a simple experiment, Wallin and Arnebrant demon-strated that a cubic phase was much faster decomposed by the action oflipase from Thermomyces (former Humicola) lanuginosa than the referencesample consisting of triolein and the aqueous phase (273). This was attrib-uted to the much larger interfacial area in the cubic phase. In an in vitrostudy of lipolysis of triglycerides in a intestinallike environment, Patton andCarey observed (37), apart from the initially occurring crystalline phase, aviscous isotropic phase composed of monoglycerides and fatty acids, whichis identical to the one formed in monoglyceride systems. In excess of bilesalts, the lipolysis products are rapidly solubilized in mixed micelles.However, the bile acid amounts in vivo are not sufficient to solubilize alllipids after a meal rich in fats, which implies that the liquid-crystallinephases exist in vivo (274). Lipase and water must be free to diffuse throughthe phases formed by the lipolysis products, surrounding the diminishing fatdroplet. Thus, the bicontinuity as well as the incorporation properties of thecubic monoglyceride phases are thought to be important featuresfor the lipolysis process (275). Recently, Borne et al., in a series of studies,have investigated the affect of lipase action on the liquid-crystalline phase aswell as other self-assemble structures such as vesicles and cubosomes(69,70,276). Some of their findings are summarized in Fig. 17, whichshows a schematic representation of the change in structure of the differentliquid crystalline phases as a function of time after adding Thermomyceslanuginose lipase. The observed changes in self-assembled structures couldbe predicted from either the mono-olein–oleic acid–aqueous ternary phasediagram, where the lipolysis give rise to a transition of cubic ! reversedhexagonal ! micellar cubic ! reversed micellar phaseþ dispersion ormono-olein–sodium oleate–aqueous ternary phase diagram, where the cor-responding sequence is lamellar ! normal hexagonal. These difference in


Figure 17 Schematic representation of the change in structure during lipolysis of

mono-olein (MO) (or diolein DO) in different liquid-crystalline phases: (a) CD phase

(63wt% MO, 37wt% 2H2O), (b) oleic acid (OA)–HII phase (65.4wt% MO, 15.6

wt% OA, 19 wt% 2H2O), (c) DO–HII phase (68wt% MO, 18wt% DO, 14wt%2H2O), and (d) La phase [10wt% MO, 5wt% sodium oleate (NaO), 85wt% 2H2O].

The main liquid-crystalline phases as determined by small-angle X-ray diffraction

(SAXD), are indicated in the figure as a diamond type of bicontinuous cubic phase,

space group Pn3m (CD), reversed hexagonal phase (HII), normal hexagonal phase

(HI), lamellar phase (La), and micellar cubic phase, space group Fd3m (Cmic). These

may exist in excess of water or in the presence of minor amounts of other phases.

Some of the observed reflections in the diffractograms, obtained by SAXD, could

not be unambiguously assigned to a structure. This unidentified structure is denoted

by X. (Adapted from Ref. 70, where details are given.)


reaction sequences could be rationalized in terms of differences in degree ofprotonation of the fatty acid (70). The initially lamellar phase had a high pH(about 10), (i.e., a low degree of protonation) and thus the degradation asexpected follows the mono-olein–sodium oleate–aqueous ternary phase dia-gram. The initially cubic and hexagonal phase had a low pH (4–7), (i.e., ahigh degree of protonation) and thus the degradation as expected followsmono-olein –oleic acid–aqueous ternary phase diagram. AddingThermomyces lanuginose lipase to aqueous dispersions of cubic phases(cubosomes) and lamellar dispersions (vesicles) at high water content andgave the corresponding morphological changes as for the liquid-crystallinephases (276). The phase diagrams of the relevant systems can thus be used asmaps to navigate through the changes in the self-assembly structure of thesubstrate and the product. Borne et al found similar specific activity ofThermomyces lanuginose lipase on the cubic as on the reversed hexagonalmono-olein-based liquid-crystalline phases, which was somewhat unex-pected (70).

ACKNOWLEDGMENTS

The fruitful discussions and collaboration with Bodil Ericsson, KareLarsson, Barry Ninham, Valdemaras Razumas, Thomas Arnebrant, BjornLindman, Ali Khan, Maura Monduzzi, Francesca Caboi, Helena Ljusberg,Marie Paulsson, Johanna Borne, Justas Barauskas, Anna Stenstam, MarieWahlgren, Fredrik Tiberg, Alan Svendsen, Karl Hult, and others aregreatly acknowledged. This work has benefited from the financial supportfrom the Swedish Research Counsil.

REFERENCES

1. P. K. J. Kinnunen, Chem. Phys. Lipids 81, 151–166 (1996).

2. P. K. J. Kinnunen and J. M. Halopainen, Biosci. Rep. 20, 465–482 (2000).

3. Y. Barenholz and G. Cevc, in Physical Chemistry of Biological Interfaces

(A. Baszkin and W. Norde, eds.), Marcel Dekker, New York, 2000, pp. 171–241.

4. H. Ringsdorf, in Physical Chemistry of Biological Interfaces (A. Baszkin and

W. Norde, eds.), Marcel Dekker, New York, 2000, pp. 243–282.

5. L. Saiz, S. Bandyopadhyay, and M. L. Klein, Biosci. Rep. 22, 151–173 (2002).

6. E. Watts, (ed.) Protein–Lipid Interactions, Elsevier, Amsterdam, 1993.

7. R. Verger and F. Pattus, Chem. Phys. Lipids 30, 189–227 (1982).

8. F. MacRitchie, Chemistry at Interfaces, Academic Press, San Diego, CA, (1990).

9. M. Ahlers, W. Muller, A. Reichert, H. Ringsdorf, and J. Venzmer, Angew.

Chem. Int. Ed. Engl. 29, 1269–1285 (1990).


10. V. B. Fainerman, E. H. Lucassen-Reynders, and R. Miller, Colloids Surfaces A:

Physicochem. Eng. Aspects 143, 141–165 (1998).

11. E. Dickinson, Collloids Surfaces B: Biointerf. 15, 161–176 (1999).

12. D. Vollhardt and V. B. Fainerman, Adv. Colloid Interf. Sci. 86, 103–151 (2000).

13. R. Miller, V. B. Fainerman, A. V. Makievski, J. Kragel, D. O. Grigoriev, V. N.

Kazakov, and O. V. Sinyachenko, Adv. Colloid Interf. Sci. 86, 39–82 (2000).

14. M. A. Bos and T. van Vliet, Adv. Colloid Interf. Sci. 91, 437–471 (2001).

15. M. J. Papiz, L. Sawyer, E. E. Eliopoulos, A. C. T. North, B. C. Findlay, R.

Sivaprasadarao, T. A. Jones, M. E. Newcomer, and P. J. Kraulis, Nature 324,

383–385 (1986).

16. L. Sawyer, Nature 327, 659 (1987).

17. A. C. T. North, Int. J. Biol. Macromol. 81, 56–58 (1989).

18. S. Brownlow, J. H. Morais Cabral, R. Cooper, D. R. Flower, S. J. Yewdall,

I. Polikarpov, A. C. T. North, and L. Sawyer, Structure 5, 481–495 (1997).

19. L. Sawyer, S. Brownlow, I. Polikarpov, and S.-Y. Wu, Int. Dairy J. 8, 65–72

(1998).

20. M. C. Diaz de Villegas, R. Oria, F. J. Salva, and M. Calvo, Milchwissenschaft

42, 357–358 (1987).

21. D. K. Sarker, P. J. Wilde, and D. C. Clark, Colloids Surfaces B: Biointerf. 3,

349–356 (1995).

22. A. Kristensen, T. Nylander, M. Paulsson, and A. Carlsson, Int. Dairy J. 7,

87–92 (1995).

23. B. Y. Qin, L. K. Creamer, E. N. Baker, and G. B. Jameson, FEBS Lett. 438,

272–278 (1998).

24. S.-Y. Wu, M. D. Perez, P. Puyol, and L. Sawyer, J. Biol. Chem. 274, 170–174

(1999).

25. L. Ragona, F. Fogolari, L. Zetta, D. M. Perez, P. Puyol, K. De Kruif, F. Lohr,

H. Ruterjans, and H. Molinari, Protein Sci. 9, 13478–1356 (2000).

26. L. Sawyer and G. Kontopidis, Biochim. Biophys. Acta 1482, 136–148 (2000).

27. L. Sawyer, P. N. Barlow, M. J. Boland, L. K. Creamer, H. Denton, P. J. B.

Edwaards, C. Holt, G. B. Jameson, G. Kontopidis, G. E. Norris, S. Uhrınova,

and S.-Y. Wu, Int. Dairy J. 12, 299–310 (2002).

28. A. L. Shapiro, E. Vinuela, and J. V. Maizel, Biochem. Biophys. Res. Commun.

28, 815–820 (1967).

29. K. Weber and M. Osborn, J. Biol. Chem. 244, 4406–4412 (1969).

30. J. Goerke, Biochim. Biophys. Acta, 1408, 79–89 (1998).

31. I. Frerking, A. Gunther, W. Seeger, and U. Pison, Intens. Care Med. 27,

1699–1717 (2001).

32. J. Perez-Gil, Biol. Neonate 81(Suppl. 1), 6–15 (2002).

33. M. Larsson, K. Larsson, S. Andersson, J. Kakhar, T. Nylander, B. Ninham,

and P. Wollmer, J. Dispers. Sci. Technol. 20, 1–12 (1999).

34. M. Larsson, K. Larsson, and P. Wollmer, Prog. Colloid Polym. Sci. 120, 28–34

(2002).

35. M. Larsson, J. J. Haitsma, B. Lachmann, K. Larsson, T. Nylander, and

P. Wollmer, Clin. Physiol. Functional Imaging 22, 39–48 (2002).


36. J. A. Clements, Proc. Soc. Exp. Biol. Med. 95, 170–172 (1957).

37. J. S. Patton and M. C. Carey, Science 204, 145–148 (1979).

38. A. Svendsen, Biochim. Biophys. Acta 1543, 223–238 (2000).

39. R. Verger, Trends Biotechnol. 15, 32–38 (1997).

40. I. Panaiotov and R. Verger, Physical Chemistry of Biological Interfaces

(A. Baszkin andW. Norde, eds.), Marcel Dekker, New York, 2000, pp. 359–400.

41. T. E. Creighton, Proteins—Structure and Molecular Properties, 2nd ed., W. H.

Freeman, New York, 1993.

42. J. S. Richardson, Adv. Protein Chem. 34, 167–339 (1981).

43. K. A. Dill, Biochemistry 29, 7133–7155 (1990).

44. T. E. Creighton, Biochem. J. 270, 1–16 (1990).

45. W. R. Taylor, A. C. W. May, N. P. Brown, and A. Aszodi, Rep. Prog. Phys.,

64, 517–590 (2001).

46. K. Larsson, Lipids—Molecular Organization, Physical Functions and Technical

Applications, Oily Press, Dundee, 1994.

47. B. Lindman and H. Wennerstrom, Topics Curr. Chem. 87, 1–83 (1980).

48. C. Tanford, The Hydrophobic Effect: Formation of Micelles and Biological

Membranes, 2nd ed., Wiley, New York, 1980.

49. K. Fontell, Adv. Colloid Interf. Sci. 41, 127–147 (1992).

50. J. N. Israelachvili, D. J. Mitchell, and B. W. Ninham, J. Chem. Soc. Faraday

Trans. II 72, 1525–1568 (1976).

51. D. J. Mitchell and B. W. Ninham, J. Chem. Soc. Faraday Trans. 2 77, 601–629

(1981).

52. A. Raudino, Adv. Colloid Interf. Sci. 57, 229–285 (1995).

53. V. Luzzati, Biological Membranes (D. Chapman, ed.), Academic Press, New

York, 1968, pp. 77–123.

54. K. Fontell, Colloid Polym. Sci. 268, 264–285 (1990).

55. R. H. Templer, Curr. Opin. Colloid Interf. Sci. 3, 255–263 (1998).

56. V. Luzzati, A. Tardieu, T. Gulik-Krzywicki, E. Rivas, and F. Reiss-Husson,

Nature 220, 485–488 (1968).

57. K. Larsson, J. Phys. Chem. 93, 7304–7314 (1989).

58. G. Lindblom and L. Rilfors, Biochim. Biophys. Acta 988, 221–256 (1989).

59. S. T. Hyde, A. S., L. K., Z. Blum, T. Landh, S. Lidin, and B. W. Ninham, The

Language of Shape. The Role of Curvature in Condensed Matter: Physics,

Chemistry and Biology, Elsevier, Amsterdam, 1997.

60. R. Wallin, S. Engstrom, and C. F. Mandenius, Biocatalysis 8, 73–80 (1993).

61. G. Lindblom, K. Larsson, L. Johansson, K. Fontell, and S. Forsen, J. Am.

Chem. Soc. 101, 5465–5470 (1979).

62. S. Andersson, S. T. Hyde, K. Larsson, and S. Lidin, Chem. Rev. 88, 221–242

(1988).

63. S. T. Hyde, J. Phys. Chem. 93, 1458–1464 (1989).

64. V. Luzzati, R. Vargas, A. Gulik, P. Mariani, J. M. Seddon, and E. Rivas,

Biochemistry 31, 279–285 (1992).

65. P. Mariani, E. Rivas, V. Luzzati, and H. Delacroix, Biochemistry 29, 6799–6810

(1990).


66. P. Mariani, V. Luzzati, and H. Delacroix, J. Mol. Biol. 204, 165–189 (1988).

67. J. Borne, T. Nylander, and A. Khan, Langmuir 17, 7742–7751 (2001).

68. J. M. Seddon, E. A. Bartle, and J. Mingins, J. Phys. Condens. Matter 2,

SA285–SA290 (1990).

69. F. Caboi, J. Borne, T. Nylander, A. Khan, A. Svendsen, and S. Patkar,

Collloids Surfaces B: Biointerf. 26, 159–171 (2002).

70. J. Borne, T. Nylander, and A. Khan, Langmuir 18, 8972–8981 (2002).

71. J. M. Seddon, Biochim. Biophys. Acta 1031, 1–69 (1990).

72. T. Landh, FEBS Lett. 369, 13–17 (1995).

73. V. Luzzati, Curr. Opin. Struct. Biol. 7, 661–668 (1997).

74. B. De Kruijff, Curr. Opin. Colloid Interf. Sci. 1, 564–569 (1997).

75. K. Larsson, Curr. Opin. Colloid Interf. Sci. 5, 64–69 (2000).

76. A. J. Verkleij, C. Mombers, W. J. Gerritsen, L. Leunissen-Bijvelt, and P. R.

Cullis, Biochim. Biophys. Acta 555, 358–361 (1979).

77. D. P. Siegel, Biophys. J. 45, 399–420 (1984).

78. J. G. Mandersloot, W. J. Gerritsen, L. Leunissen-Bijvelt, A. Van Echteld, P. C.

Noordam, and J. De Gier, Biochim. Biophys. Acta 640, 106–113 (1981).

79. M. J. Hope and P. R. Cullis, Biochim. Biophys. Acta 640, 82–90 (1981).

80. W. Helfrich, Liquid Crystallogr. 5, 1647–1658 (1989).

81. D. D. Lasic, Liposomes—From Physics to Applications, Elsevier, Amsterdam,

1993.

82. S. Komura, in Vesicles (M. Rosoff, ed.), Marcel Dekker, New York, 1996,

pp. 198–236.

83. D. D. Lasic, R. Joannic, B. C. Keller, P. M. Frederik, and L. Auvray, Adv.

Colloid Interf. Sci. 89–90, 337–349 (2001).

84. T. Landh, J. Phys. Chem. 98, 8453–8467 (1994).

85. J. Gustafsson, H. Ljusberg-Wahren, M. Almgren, and K. Larsson, Langmuir

12, 4611 (1996).

86. J. Gustafsson, H. Ljusberg-Wahren, M. Almgren, and K. Larsson, Langmuir

13, 6964 (1997).

87. C. Tanford, Physical Chemistry of Macromolecules, Wiley, New York, 1967.

88. K. Gekko and Y. Hasegawa, Biochemistry 25, 6563–6571 (1986).

89. P. L. Privalov, Adv. Protein Chem. 33, 167–241 (1979).

90. P. L. Privalov, Adv. Protein Chem. 35, 1–104 (1982).

91. P. L. Privalov and S. J. Gill, Adv. Protein Chem. 39, 191–234 (1988).

92. P. K. Ponnuswamy, Prog. Biophys. Mol. Biol. 59, 57–103 (1993).

93. C. N. Pace, L. M. Fisher, and J. F. Cupo, Acta Biol. Med. Germ. 40, 1385–1392

(1981).

94. F. M. Richards, Annu. Rev. Biophys. Bioeng. 6, 151–176 (1977).

95. W. Kauzmann, Adv. Protein Chem. 14, 1–63 (1959).

96. W. Norde, Adv. Colloid Interf. Sci. 25, 267–340 (1986).

97. C. A. Haynes and W. Norde, Colloids Surfaces B: Biointerf. 2, 517–566

(1994).

98. W. Norde, Physical Chemistry of Biological Interfaces (A. Baszkin and

W. Norde, eds.), Marcel Dekker, New York, 2000, pp. 115–135.


99. M. C. L. Maste, W. Norde, and A. J. W. G. Visser, J. Colloid Interf. Sci. 196,

224–230 (1997).

100. T. Zoungrana, G. H. Findenegg, and W. Norde, J. Colloid Interf. Sci. 190,

437–448 (1997).

101. D. A. Dolgikh, R. I. Gilmanshin, E. V. Brazhnikov, V. E. Bychkova, G. V.

Semisotnov, S. Y. Venyaminov, and O. B. Ptitsyn, FEBS Lett. 136, 311–315

(1981).

102. M. Ohgushi and A. Wada, FEBS Lett. 164, 21–24 (1983).

103. D. A. Dolgikh, L. V. Abaturov, I. A. Bolotina, E. V. Brazhnikov, V. E.

Bychkova, R. I. Gilmanshin, Y. O. Lebedev, G. V. Semisotnov, E. I.

Tiktopulo, and O. B. Ptitsyn, Eur. Biophys. J. 13, 109–121 (1985).

104. K. Kuwajima, Proteins: Struct. Funct. Genet. 6, 87–103 (1989).

105. O. B. Ptitsyn, R. H. Pain, G. V. Semisotnov, E. Zerovnik, and O. I.

Razgulyaev, FEBS Lett. 262, 20–24 (1990).

106. E. Dickinson and Y. Matsumura, Colloids Surfaces B: Biointerf. 3, 1–17

(1994).

107. V. E. Bychokova, R. H. Pain, and O. B. Ptitsyn, FEBS Lett. 238, 231–234

(1988).

108. F. G. van der Goot, J. M. Gonzalez-Manas, J. H. Lakey, and F. Pattus,

Nature 354, 408–410 (1991).

109. M. Bostrom, D. R. M. Williams, and B. W. Ninham, Phys. Rev. Lett. 87,

168103 (2001).

110. M. Bostrom, D. R. M. Williams, and B. Ninham, Langmuir 18, 8609–8615

(2002).

111. B. W. Ninham, Progr. Colloid Polym. Sci. 120, 1–12 (2002).

112. J. Steinhardt and J. A. Reynolds, Multiple Equilibria in Proteins, Academic

Press, New York, 1969.

113. S. Lapanje, Physicochemical Aspects of Protein Denaturation, Wiley, New

York, 1978.

114. S. Makino, Adv. Biophys. 12, 131–184 (1979).

115. M. N. Jones and A. Brass, Food, Polymers, Gels, and Colloids (E. Dickinson,


116. K. P. Ananthapadmanabhan, Interactions of Surfactants with Polymers and

Proteins (K. P. Ananthapadmanabhan and G. E. D., eds.), CRC Press,

Boca Raton, FL, 1993, pp. 319–366.

117. E. Dickinson, Interactions of Surfactants with Polymers and Proteins (K. P.

Ananthapadmanabhan and G. E. D., eds.), CRC Press, Boca Raton, FL,

1993, pp. 295–317.

118. M. Bos, T. Nylander, T. Arnebrant, and D. C. Clark, Food Emulsifiers and

Their Applications (G. L. Hasenhuettl and R. W. Hartel, eds.), Chapman &

Hall, New York, 1997, pp. 95–146.

119. M. N. Jones and P. Manley, J. Chem. Soc. Faraday Trans. 1 75, 1736–1744

(1979).

120. M. N. Jones and P. Manley, J. Chem. Soc. Faraday Trans. 1 76, 654–664

(1980).


121. A. Yonath, A. Podjarny, H. B., A. Sielecki, and W. Traub, Biochemistry 16,

1418–1424 (1977).

122. A. Yonath, A. Sielecki, J. Moult, A. Podjarny, and W. Traub, Biochemistry

16, 1413–1417 (1977).

123. R. Waninge, M. Paulsson, T. Nylander, B. Ninham, and P. Sellers, Int. Dairy

J. 8, 141–148 (1998).

124. J. A. Reynolds and C. Tanford, Proc. Natl. Acad. Sci. USA 66, 1002–1007

(1970).

125. R. Miller and K. Lunkenheimer, Colloid Polym. Sci. 264, 273–276 (1986).

126. K. Lunkenheimer and R. Miller, J. Colloid Interf. Sci. 120, 176–183 (1987).

127. K. Lunkenheimer and G. Czichocki, J. Colloid Interf. Sci. 160, 509–510 (1993).

128. D. C. Clark, F. Husband, P. J. Wilde, M. Cornec, R. Miller, J. Kragel, and

R. Wustneck, J. Chem. Soc. Faraday Trans. 91, 1991–1996 (1995).

129. P.-O. Hegg, Acta Agric. Scand. 30, 401–404 (1980).

130. S. Gumpen, P.-O. Hegg, and H. Martens, Biochim. Biophys. Acta 574,

189–196 (1979).

131. P. Puyol, M. D. Perez, J. M. Peiro, and M. Calvo, J. Dairy Sci. 77, 1494–1502

(1994).

132. K. Fukushima, Y. Murata, N. Nishikido, G. Sugihara, and M. Tanaka, Bull.

Chem. Soc. Jpn. 54, 3122–3127 (1981).

133. K. Fukushima, Y. Murata, G. Sugihara, and M. Tanaka, Bull. Chem. Soc.

Jpn. 55, 1376–1378 (1982).

134. J. Chen and E. Dickinson, Colloids Surfaces A: Physicochem. Eng. Aspects

100, 255–265 (1995).


100, 267–277 (1995).

136. A. K. Moren and A. Khan, Langmuir 11, 3636–3643 (1995).

137. A. Stenstam, A. Khan, and H. Wennerstrom, Langmuir 17, 7513–7520 (2001).

138. W. L. Mattice, J. M. Riser, and D. S. Clark, Biochemistry 15, 4264–4272

(1976).

139. B. Jirgensons, Biochim. Biophys. Acta 434, 58–68 (1976).

140. R. E. Tanner, B. Herpigny, S. H. Chen, and C. K. Rha, J. Chem. Phys. 76,

3866–3872 (1982).

141. X. H. Guo and S. H. Chen, Chem. Phys. 149, 129–139 (1990).

142. X. H. Guo, N. M. Zhao, S. H. Chen, and J. Teixeira, Biopolymers 29, 335–346

(1990).

143. C. A. Nelson, J. Biol. Chem. 246, 3895–3901 (1971).

144. M. N. Jones, P. Manley, P. J. W. Midgley, and A. E. Wilkinson, Biopolymers

21, 1435–1450 (1982).

145. A. K. Moren and A. Khan, Langmuir 14, 6818–6826 (1998).

146. J. Reynolds, S. Herbert, and J. Steinhardt, Biochemistry 7, 1357–1361 (1968).

147. H. M. Rendall, J. Chem. Soc. Faraday Trans. 1, 72, 481–484 (1976).

148. S. Makino, J. A. Reynolds, and C. Tanford, J. Biol. Chem. 248, 4926–4932

(1973).

149. F. A. Green, J. Colloid Interf. Sci. 35, 481–485 (1971).


150. J. Cordoba, M. D. Reboiras, and M. N. Jones, Int. J. Biol. Macromol. 10,

270–276 (1988).

151. W. W. Sukow, H. E. Sandberg, E. A. Lewis, D. J. Eatough, and L. D. Hansen,

Biochemistry 19, 912–917 (1980).

152. P. J. Wilde and D. C. Clark, J. Colloid Interf. Sci. 155, 48–54 (1993).

153. D. C. Clark, P. J. Wilde, D. R. Wilson, and R. Wustneck, Food Hydrocolloids

6, 173–186 (1992).

154. P. J. Wilde, D. C. Clark, and D. Marion, J. Agric. Food Chem. 41, 1570–1576

(1993).

155. C. Tanford and Epstein, J. Am. Chem. Soc. 76, 2163–2169 (1954).

156. S. Kaneshina, M. Tanaka, T. Kondo, T. Mizuno, and K. Aoki, Bull. Chem.

Soc. Jpn. 46, 2735–2738 (1973).

157. Y. Nozaki, J. A. Reynolds, and C. Tanford, J. Biol. Chem. 249, 4452–4459

(1974).

158. B. Ericsson, P.-O. Hegg, and K. Martensson, J. Dispers. Sci. Technol. 8,

271–287 (1987).

159. B. Ericsson, P.-O. Hegg, and K. Martensson, J. Dispers. Sci. Technol. 8,

289–301 (1987).

160. M. N. Jones, P. Manley, and A. Holt, Int. J. Biol. Macromol. 6, 65–68 (1984).

161. R. J. Williams, J. N. Phillips, and K. J. Mysels, Trans. Faraday Soc. 51,

728–737 (1955).

162. B. D. Flockhart, J. Colloid Interf. Sci. 16, 484–492 (1961).

163. D. F. Evans, D. J. Mitchell, and B. W. Ninham, J. Phys. Chem. 88, 6344–6348

(1984).

164. D. F. Evans, M. Allen, B. W. Ninham, and A. Fouda, J. Solution Chem. 13,

87–101 (1984).

165. B. Ericsson, P.-O. Hegg, and K. Martensson, J. Food Technol. 18, 11–19

(1983).

166. J. A. Reynolds, J. P. Gallagher, and J. Steinhardt, Biochemistry 9, 1232–1238

(1970).

167. M. Subramanian, B. S. Sheshadri, and M. P. Venkatappa, J. Biochem.

(Tokyo) 95, 413–421 (1984).

168. D. C. Clark, M. Coke, P. J. Wilde, and D. R. Wilson, Food, Polymers, Gels

and Colloids (E. Dickinson, ed.), Royal Society of Chemistry, London, 1991,

pp. 272–278.

169. W. E. Ewers and K. L. Sutherland, Austr. J. Sci. Res. Ser. A5, 697–710 (1952).

170. N. Nishikido, T. Takahara, H. Kobayashi, and M. Tanaka, Bull. Chem. Soc.

Jpn. 55, 3085–3088 (1982).

171. B. Ericsson and P.-O. Hegg, Prog. Colloid Polym. Sci. 70, 92–95 (1985).

172. M. Wahlgren, S. Welin-Klintstrom, T. Arnebrant, A. Askendal, and H.

Elwing, Colloids Surfaces B: Biointerf. 4, 23–31 (1995).

173. A. R. Mackie, A. P. Gunning, P. J. Wilde, and V. J. Morris, J. Colloid Interf.

Sci. 210, 157–166 (1999).

174. A. R. Mackie, A. P. Gunning, M. J. Ridout, P. J. Wilde, and V. J. Morris,

Langmuir 17, 6593–6598 (2001).


175. A. R. Mackie, A. P. Gunning, M. J. Ridout, P. J. Wilde, and J. Rodriguez

Patino, Biomacromolecules 2, 1001–1006 (2001).

176. J. R. Lu, T. J. Su, and R. K. Thomas, J. Phys. Chem. B 102, 10,307–10,315

(1998).

177. J. Chen and E. Dickinson, J. Sci. Food Agric. 62, 283–289 (1993).

178. J. Chen, E. Dickinson, and G. Iveson, Food Struc. 12, 135–146 (1993).

179. E. Dickinson and Y. Matsumura, Int. J. Biol. Macromol. 13, 26–30 (1991).

180. J. McGuire, M. C. Wahlgren, and T. Arnebrant, J. Colloid Interf. Sci. 170,

182–192 (1995).

181. M. C. Wahlgren and T. Arnebrant, J. Colloid Interf. Sci. 142, 503–511

(1991).

182. R. J. Green, T. J. Su, J. R. Lu, and J. Penfold, J. Phys. Chem. B 105,

1594–1602 (2001).

183. R. J. Green, T. J. Su, H. Joy, and J. R. Lu, Langmuir 16, 5797–5805

(2000).

184. M. C. Wahlgren and T. Arnebrant, J. Colloid Interf. Sci. 148, 201–206

(1992).

185. B. Ericsson, Interactions Between Globular Proteins and Lipids, PhD Thesis,

University of Lund, Lund, Sweden, 1986.

186. R. Miller, V. B. Fainerman, A. V. Makievski, J. Kragel, and R. Wustneck,

Colloids Surfaces A: Physicochem. Eng. Aspects 161, 151–157 (2000).

187. J. J. Garcia Dominguez, R. Infante, P. Erra, and R. Julia, Tenside Detergents

18, 310–313 (1981).

188. J. H. Buckingham, J. Lucassen, and D. Giles, J. Colloid Interf. Sci. 67,

423–431 (1978).

189. E. H. Lucassen-Reynders, J. Lucassen, and D. Giles, J. Colloid Interf. Sci. 81,

150–157 (1981).

190. A. Prins and D. J. M. Bergink-Martens, Food Colloids and Polymers: Stability

and Mechanical Properties (E. Dickinson and P. Walstra, eds.), The Royal


191. E. Dickinson and G. Stainsby, Colloids in Food, Applied Science Publishers,

London, 1982.

192. M. Coke, P. J. Wilde, E. J. Russell, and D. C. Clark, J. Colloid Interf. Sci. 138,

489–504 (1990).

193. J. Kragel, R. Wustneck, D. Clark, P. Wilde, and R. Miller, Colloids Surfaces

A: Physicochem. Eng. Aspects 98, 127–135 (1995).


101, 77–85 (1995).

195. M. A. Bos and T. Nylander, 12, 2791–2797 (1996).

196. S. G. Hambling, A. S. McAlpine, and L. Sawyer, Advanced Dairy Chemistry.

Vol. 1: Proteins (P. Fox, ed.), Elsevier Applied Science, London, 1992,

pp. 141–190.

197. D. G. Cornell and D. L. Patterson, J. Agric. Food Chem. 37, 1455–1459

(1989).

198. D. G. Cornell, J. Colloid Interf. Sci. 88, 536–545 (1982).


199. D. G. Cornell, D. L. Patterson, and N. Hoban, J. Colloid Interf. Sci. 140,

428–435 (1990).

200. P. J. Quinn and R. M. C. Dawson, Biochem. J. 113, 791–803 (1969).

201. P. J. Quinn and R. M. C. Dawson, Biochem. J. 115, 65–75 (1969).

202. M. Malmsten, P. Claesson, and G. Siegel, Langmuir 10, 1274–1280 (1994).

203. M. Malmsten, J. Colloid Interf. Sci. 172, 106–115 (1995).

204. Y. K. Du, J. Y. An, J. Tang, Y. Li, and L. Jiang, Colloid Surfaces B: Biointerf.

7, 129–133 (1996).

205. J. A. Ibdah and M. C. Phillips, Biochemistry 27, 7155–7162 (1988).

206. I. Hanssens and F. H. Van Cauwelaert, Biochem. Biophys. Res. Commun. 84,

1088–1096 (1978).

207. L. Lindahl and H. J. Vogel, Anal. Biochem. 140, 394 (1984).

208. W. M. Heckl, B. N. Zaba, and H. Mohwald, Biochim. Biophys. Acta 903,

166–176 (1987).

209. M. Schonhoff, M. Losche, M. Meyer, and C. Wilhelm, Prog. Colloid Polym.

Sci. 89, 243–248 (1992).

210. J. Zhao, D. Vollhardt, G. Brezesinski, S. Siegel, J. Wu, J. B. Li, and R. Miller,


211. D. G. Cornell and R. J. Caroll, J. Colloid Interf. Sci. 108, 226–233 (1985).

212. A. Diederich, C. Sponer, D. Pum, U. B. Sleytr, and M. Losche, Colloids

Surfaces B: Biointerf. 6, 335–346 (1996).

213. D. Vollhardt, Adv. Colloid Interf. Sci. 47, 1–23 (1993).

214. D. Vollhardt, T. Kato, and M. Kawano, J. Phys. Chem. 100, 4141–4147 (1996).

215. A. Gericke, J. Simon-Kutscher, and H. Huhnerfuss, Langmuir 9, 2119–2127

(1993).

216. A. Carlsson, M. Bergqvist, R. Nilsson, and T. Nylander, Program Drug

Delivery Systems 5, 105-115 (1995).

217. S. Danithine, C. Blecker, M. Paquot, N. Innocente, and C. Deroanne, Lait 80,

209–222 (2000).

218. I. H. Mather, J. Dairy Sci. 83, 203–247 (2000).

219. D. Kristensen, T. Nylander, J. T. Rasmussen, M. Paulsson, and A. Carlsson,

Langmuir 12, 5856–5862 (1996).

220. D. J. Siminovitch and K. R. Jeffrey, Biochim. Biophys. Acta 645, 270–278

(1981).

221. Y. Yamamoto and M. Araki, Biosci. Biotech. Biochem. 61, 1791–1795 (1997).

222. S. Aynie, M. Le Meste, B. Colas, and D. Lorient, J. Food Sci. 57, 883–887

(1992).

223. E. Bylaite, T. Nylander, R. Venskutonis, and B. Jonsson, Collloids Surfaces B:

Biointerf. 20, 327–340 (2001).

224. J. L. Courthaudon, E. Dickinson, and W. W. Christie, J. Agric. Food Chem.

39, 1365–1368 (1991).

225. E. Dickinson and G. Iveson, Food Hydrocolloids 6, 533–541 (1993).

226. Y. Fang and D. G. Dalgleish, J. Am. Oil Chem. Soc. 73, 437–442 (1996).

227. I. Heertje, J. Nederlof, H. A. C. M. Hendrickx, and E. H. Lucassen-Reynders,

Food Struct. 9, 305–316 (1990).


228. U. Seifert, K. Berndl, and R. Lipowsky, Phys. Rev. A 44, 1182–1202 (1991).

229. R. Waninge, T. Nylander, M. Paulsson, and B. Bergenstahl, Collloids Surfaces

B: Biointerf. in press (2003).

230. M. Rytomaa, P. Mustonen, and P. K. J. Kinnunen, J. Biol. Chem. 267,

22,243–22,248 (1992).

231. M. E. Price, R. M. Cornelius, and J. L. Brash, Biochim. Biophys. Acta 1512,

191–205 (2001).

232. D. V. Brooksbank, J. Leaver, and D. S. Horne, J. Colloid Interf. Sci. 161,

38–42 (1993).

233. R. W. Corkery, Collloids Surfaces B: Biointerf. 26, 3–20 (2002).

234. J. Kim and H. Kim, Biochemistry 25, 7867–7874 (1986).

235. E. M. Brown, R. J. Caroll, P. E. Pfeffer, and J. Sampugna, Lipids 18, 111–118

(1983).

236. E. M. Brown, J. Dairy Sci. 67, 713–722 (1984).

237. J. N. de Wit, Developments in Dairy Chemistry—4 (P. F. Fox, ed.), Elsevier

Applied Science, London, 1989, pp. 285–322.

238. A. Raudino and F. Castelli, Colloid Polym. Sci. 270, 1116–1123 (1992).

239. T. Hønger, K. Jørgensen, R. L. Biltonen, and O. G. Mouritsen, Biochemistry

35, 9003–9006 (1996).

240. H. Minami, T. Nylander, and A. Carlsson, Phys. Chem. Lipids, submitted

(1995).

241. D. M. LeNeveu, R. P. Rand, and V. A. Parsegian, Biophys. J. 18, 209–230

(1977).

242. A. C. Cowley, N. L. Fuller, R. P. Rand, and V. A. Parsegian, Biochemistry 17,

3163–3168 (1978).

243. R. P. Rand, Biochim. Biophys. Acta 241, 823–834 (1971).

244. C. D. McCallum and R. M. Epand, Biochemistry 34, 1815–1824 (1995).

245. B. A. Wallis, Biophys. J. 49, 295–306 (1986).

246. V. Chupin, J. A. Killian, and B. De Kruijff, Biophys. J. 51, 395–405 (1987).

247. B. De Kruijff and P. R. Cullis, Biochim. Biophys. Acta 602, 477–490 (1980).

248. P. J. R. Spooner and A. Watts, Biochemistry 30, 3880–3885 (1991).

249. P. J. R. Spooner and A. Watts, Biochemistry 30, 3871–3879 (1991).

250. T. Heimburg, P. Hildebrandt, and D. Marsh, Biochemistry 30, 9084–9089

(1991).

251. V. Razumas, K. Larsson, Y. Miezes, and T. Nylander, J. Phys. Chem. 100,

11,766–11,774 (1996).

252. J. Barauskas, V. Razumas, and T. Nylander, Prog. Colloid Polym. Sci. 116,

16–20 (2000).

253. P. E. Fraser, R. P. Rand, and C. M. Deber, Biochim. Biophys. Acta 983, 23–29

(1989).

254. K. Larsson, Nature 304, 664 (1983).

255. S. T. Hyde, S. Andersson, B. Ericsson, and K. Larsson, Z. Kristallogr. 168,

213–219 (1984).

256. J. Briggs, H. Chung, and M. Caffrey, J. Phys. II France 6, 723–751 (1996).

257. H. Qui and M. Caffrey, Biomaterials 21, 223–234 (2000).


258. C. Mattisson, T. Nylander, A. Axelsson, and G. Zacchi, Chem. Phys. Lipids

84, 1–12 (1996).

259. V. Razumas, J. Kanapieniene, T. Nylander, S. Engstrom, and K. Larsson,

Anal. Chim. Acta 289, 155–162 (1994).

260. V. Razumas, Z. Talaikyte, J. Barauskas, K. Larsson, Y. Miezis, and T.

Nylander, Chem. Phys. Lipids 84, 123–138 (1996).

261. T. Nylander, C. Mattisson, V. Razumas, Y. Miezis, and B. Hakansson,


262. S. B. Leslie, S. Puvvada, B. R. Ratna, and A. S. Rudolph, Biochim. Biophys.

Acta 1285, 246–254 (1996).

263. B. Ericsson, K. Larsson, and K. Fontell, Biochim. Biophys. Acta 729, 23–27

(1983).

264. M. Portmann, E. M. Landau, and P. L. Luisi, J. Phys. Chem. 95, 8437–8440

(1991).

265. E. M. Landau and P. L. Luisi, J. Am. Chem. Soc. 115, 2102–2106 (1993).

266. K. Larsson and G. Lindblom, J. Dispers. Sci. Technol. 3, 61–66 (1982).

267. E. M. Landau and J. P. Rosenbusch, Proc. Natl. Acad. Sci. USA 93, 14532–

14535 (1996).

268. J. Baruskas, V. Razumas, and T. Nylander, Chem. Phys. Lipids 97, 167–179

(1999).

269. H. Gutman, G. Arvidson, K. Fontell, and G. Lindblom, Surfactants in

Solution (K. L. Mittal and B. Lindman, eds.), Plenum, New York, 1984, pp.

143–152.

270. J. Engblom, Y. Miezis, T. Nylander, V. Razumas, and K. Larsson, Prog.

Colloid Polym. Sci. 116, 9–15 (2000).

271. F. Caboi, T. Nylander, V. Razumas, Z. Talaikyte, M. Monduzzi, and K.

Larsson, Langmuir, 13, 5476–5483.

272. F. Caboi, G. S. Amico, P. Pitzalis, M. Monduzzi, T. Nylander, and K.

Larsson, Chem. Phys. Lipids 109, 47–62 (2001).

273. R. Wallin and T. Arnebrant, J. Colloid Interf. Sci. 164, 16–20 (1994).

274. M. Lindstrom, H. Ljusberg-Wahren, K. Larsson, and B. Borgstrom, Lipids

16, 749–754 (1981).

275. J. S. Patton, R. D. Vetter, M. Hamosh, B. Borgstrom, M. Lindstrom, and

M. C. Carey, Food Microstruct. 4, 29–41 (1985).

276. J. Borne, T. Nylander, and A. Khan, J. Phys. Chem. B 106, 10,492–10,500

(2002).

277. K. Kurihara and Y. Katsuragi, Nature 365, 213–214 (1993)


5Droplet Flocculation andCoalescence in DiluteOil-in-Water Emulsions

Øystein Sæther and Johan SjoblomNorwegian University of Science and Technology, Trondheim, Norway

Stanislav S. DukhinNew Jersey Institute of Technology, Newark, New Jersey, U.S.A.

I. INTRODUCTION

Emulsions are common in the food industry. They are presented to theconsumer as finished products or they appear during the preparation offood from the mixing and treatment of suitable constituents. The emulsionhas a certain texture which influences the perception of the food aestheticsand which to the consumer is the most important feature. At the basis liesemulsion stability, which represents the scientific point of view. The texturecan be interpreted as a function of the physicochemical properties of theemulsion—the dispersed concentration, droplet size and interactions, bulkand interfacial rheology, and so on—properties that also control dropletaggregation (flocculation or coagulation) and rupture of the membranebetween aggregated droplets (coalescence).

Emulsions are thermodynamically unstable (unlike microemulsions),and stability is achieved kinetically. The notion of stability includes bothretardation of sedimentation or creaming (driven by the density differencebetween droplets and the surrounding fluid, and greatly dependent onthe viscosity of the last and droplet size), reduction of coagulation [irre-versible droplet aggregation, countered (e.g., by repulsion arising from the


adsorbed layer)], and stabilization of the interface in order to retardcoalescence (linked to the viscoelastic properties of the interfacial layer).

Food emulsions include milk, butter, ice cream, mayonnaise,margarine, and many more. Milk, cream, and mayonnaise are oil-in-wateremulsions, stabilized from coalescence by interfacially adsorbed surfactantsindigenous to the raw material.

The next example illustrates how the processes of destabilization arecountered. The homogenization of milk induces droplet breakup by passingthe liquid through a high-pressure-drop mixing valve, resulting in dropletsizes down to about 250 nm (1). A droplet size of 250 nm is so small that thecreaming rate is extremely low and Brownian diffusion will dominate drop-let motion. The newly formed interface is stabilized by adsorbed casein (andrelatives) and phospholipids, expectedly forming a complex film of mono-mers, aggregates, and even particles (2). Adsorbed species at the interfacegive rise to a repulsive contribution to the droplet–droplet interaction poten-tial (electrostatic and/or steric), reducing the collision efficiency (the ratio ofcollisions leading to coalescence to the total number of collisions). The filmformed by the species can, to some system-specific degree, resist ruptureupon collision, as elastic film properties will restore the film as collisionenergy is dissipated. These separate factors work to decrease the rate ofseparation and give milk a certain lifetime. Figure 1 shows snapshots of

Figure 1 Image series (from frame 1 at the upper left to frame 12 at the lower right)

showing oil droplets in milk interacting, and the formation and fragmentation of a

short-lived doublet. The milk (‘‘1.5% fat’’) was centrifuged to reduce the droplet

numerical concentration in the prepared sample. The largest droplet is about 3 mm,

the time between images 0.5 s.


droplets interacting in a milk sample in which the majority of droplets havebeen centrifuged off to isolate droplet pair effects.

In the following, the authors attempt to describe the interplayof mechanisms that influence the breakdown of kinetically stabilizedemulsions, namely droplet aggregation (flocculation or coagulation),droplet aggregate fragmentation and coalescence.

A. Kinetic and Thermodynamic Stability inMacroemulsions and Miniemulsions

The majority of emulsion technology problems relate to the stabilizationand destabilization of emulsions (3–9). Despite the existence of many fun-damental studies related to the stability of emulsions, the extreme vari-ability and complexity of the systems involved in any specific applicationoften pushes the industry to achieve technologically applicable resultswithout developing a detailed understanding of the fundamental processes.Nevertheless, because in most cases, technological success requires thedesign of emulsions with a very delicate equilibrium between stabilityand instability, a better understanding of the mechanisms of stabiliza-tion and destabilization might lead to significant breakthroughs intechnology.

Notwithstanding their thermodynamic instability, many emulsions arekinetically stable and do not change appreciably for a prolonged period.These systems exist in the metastable state (10–17). The fundamentals ofemulsion stability (destabilization) comprise emulsion surface chemistry andphysicochemical kinetics.

In contrast to the large success in industrial applications of emulsionsurface chemistry, the potential of physicochemical kinetics as basis foremulsion dynamics modeling is almost never used in emulsion technology.This situation has started to change during the last decade. Althoughthe coupling of the subprocesses in emulsion dynamics modeling (EDM)continues to represent a large problem yet to be solved, models areelaborated for (a) macroemulsions (12,18–24) and (b) miniemulsions(25–32) for long and short lifetimes of thin emulsion films.

(a) For large droplets (larger than 10–30 mm) in macroemulsions, therate of thinning of the emulsion film formed between twoapproaching droplets is rather low and, correspondingly, theentire lifetime of an emulsion need not be short, even withoutsurfactant stabilization of the film. For this case, the notion ofkinetic stability is introduced (12,18–21) to denote the resistanceof the film against rupture during thinning. Droplet deformation


and flattening of the interface is the cause of this strongresistance, described by the Reynolds equation (33,34).According to theory, the role of deformation (35–37) decreasesrapidly with decreasing droplet dimension.

(b) For small droplets (smaller than 5–10 mm) in miniemulsions,droplet deformation can be neglected, because the Reynoldsdrainage rate increases as (12,38) (Rd ¼ the Reynolds film radius)and because the smaller the droplets, the smaller is thedeformation (35–37).

In distinction from macroemulsions, where the kinetic stability is themanifestation of droplet–droplet hydrodynamic interaction and dropletdeformation, in miniemulsions the kinetic stability is the manifestationof the interplay between surface forces and Brownian movement (25).As the molecular forces of attraction decrease linearly with decreasingdroplet dimension (viz. approximately 10 times at the transitionfrom macroemulsions to miniemulsions) the potential minimum ofdroplet–droplet interaction (the secondary minimum) decreases, and forminiemulsions, this depth can be evaluated as 1–5 kT (14,39). At suchlow energy, Brownian movement causes disaggregation of droplet doubletsafter a short time (the doublet fragmentation time, �d ). If this timeis shorter than the lifetime of the thin film, a rapid decrease inthe total droplet concentration (t.d.c) is prevented (restricted by thecoalescence time, �c); that is, stability is achieved due to this kineticmechanism (25).

B. Current State of Emulsion Stability Science

A large misbalance exists between knowledge concerning kinetic stabilityand thermodynamic stability. Attention has been paid primarily to kineticstability for both macroemulsions (18–24) and miniemulsions (25–32).As a result, the droplet–droplet interaction and the collective processesin dilute emulsions are quantified (40,41) and important experimentalinvestigations have been made (29,30,42). Some models are elaboratedfor the entire process of coalescence in concentrated emulsions as well(43,44). Given thermodynamic stability, a thin interdroplet film can bemetastable.

In contrast to the large achievements in the investigations of kineticstability, modest attention has been paid to the fundamentals of the thermo-dynamic stability in emulsions, especially regarding the surfactantadsorption layer’s influence on the coalescence time. There are severalinvestigations devoted to the surface chemistry of adsorption related to


emulsification and demulsification. However, the link among the chemicalnature of an adsorption layer, its structure and the coalescence time is notyet quantified.

A premise for such quantification is the theory of a foam bilayerlifetime (45). The main notions of this theory is similar to theory ofDerjaguin et al. (46,47). However, the theory (45) is specified for amphiphilefoam films, it is elaborated in detail, and it is proven by experiment withwater-soluble amphiphiles, such as sodium dodecyl sulfate (48). Becauseas the dependence of the rupture of the emulsion film on surfactantconcentration is similar to that of a foam film, modification of the theorywith respect to emulsions may be possible. Although this modification isdesirable, the specification of theory for a given surfactant will not be trivial,because the parameters in the equation for the lifetime (47) are unknownand their determination is not trivial. As the theory (45,49) is proposedfor amphiphiles and a wider class of chemical compounds can stabilizeemulsions, the film rupture mechanism (46) is not universal regardingemulsions.

In contrast to the quantification of kinetic stability, the empiricalapproach continues to predominate regarding thermodynamic stability.Meanwhile, thermodynamic stability provides greater opportunity forlong-term stabilization of emulsions than does kinetic stability. Thismeans that the experimental characterization of thermodynamic stability(i.e., the measurement of coalescence time) is of major importance.

C. Specificity of Emulsion Characterization

Generalized emulsion characterization (i.e., measurement of droplet sizedistribution, electrokinetic potential, Hamaker constant, etc.) is notalways sufficient. Thermodynamic stability with respect to bilayer rupturecannot be quantified with such a characterization procedure alone.Consequently, measurement of the coalescence time �c is of major impor-tance for an evaluation of emulsion stability; it is an important and specificparameter of emulsion characterization.

The current state of miniemulsion characterization neglects the impor-tance of �c measurement. The practice of �c measurement is practically absentwith the exception of only a few articles considered in this chapter.Meanwhile, many articles devoted to issues more or less related to emulsionstability do not discuss �c measurement. One reason for this scientifically andtechnologically unfavorable situation in which emulsions are incompletelycharacterized might originate from a lack of devices enabling �cmeasurement.


D. Scope of the Chapter

This chapter is focused on kinetic stability in miniemulsions, with emphasison the coupled destabilizing subprocesses. In general, there are threecoupled subprocesses which will influence the rate of destabilization andphase separation in emulsions. These are aggregation, coalescence, andfloc fragmentation. Often, irreversible aggregation is called coagulationand the term flocculation is used for reversible aggregation (45,50).Ostwald ripening (51,52) coupled (26) with aggregation and fragmentationis a separate topic and will be not considered here.

A simplified theory is available for the coupling of coalescenceand flocculation in emulsions void of larger flocs. This theory is consid-ered in Section II and will assist in the consideration of the more compli-cated theory of coupling of coalescence and coagulation (Sec. III). Theexperimental investigations are described in parallel. Section IV is devotedto the theory of doublet fragmentation time and its measurement, asthis characterizes an emulsion regarding fragmentation and because itsmeasurement is an important source of information about surface forcesand the pair interaction potential. The discrimination between conditionsfor coupling of coalescence with coagulation or with flocculation isconsidered in Section V. The quantification of kinetic stability createsnew opportunities for the long-term prediction of miniemulsion stability,for stability optimization, and for characterization with standardizationof �c and �d measurements. This forms the base for emulsion dynamicsmodeling (Sec. VI).

II. COUPLING OF COALESCENCE AND FLOCCULATION

A. Singlet–Doublet Quasiequilibrium

Each process among the three processes under consideration is charac-terized by a characteristic time, namely �Sm, �d , and �c. TheSmoluchowski time (53), �Sm, gives the average time between droplet colli-sions. If the time between two collisions is shorter than �d , a doublet cantransform into a triplet before it spontaneously disrupts. In the oppositecase—at

�Sm >> �d ð1Þ

—the probability for a doublet to transform into a triplet is very lowbecause the disruption of the doublet occurs much earlier than its collision


with a singlet. The rate of multiplet formation is very low for

Rev ¼�d�Sm

<< 1 ð2Þ

where we introduce the notation Rev for small values of the ratiocorresponding to the reversibility of aggregation and a singlet–doubletquasiequilibrium.

The kinetic equation for reversible flocculation in a dilute mono-disperse oil-in-water (o/w) emulsion when neglecting coalescence is (54–56)

dn2

dt¼

n21�Sm

�n2

�dð3Þ

where n1 and n2 are the dimensionless concentrations of doublets andsinglets and n1 ¼ N1=N10 and n2 ¼ N2=N10, where N1 and N2 are theconcentrations of singlets and doublets, respectively, and N10 is the initialconcentration, and

�Sm ¼4kT

3�N10

� ��1

¼ Kf N10

� ��1ð4Þ

where k is the Boltzmann constant, T is the absolute temperature, and� is the viscosity of water. For aqueous dispersions at 25�C, Kf ¼

ð4kT=3�Þ ¼ 6� 10�18 m3=s. The singlet concentration decreases with timedue to doublet formation, whereas the doublet concentration increases. As aresult, the rates of aggregation and floc fragmentation will approach eachother. Correspondingly, the change in the number of doublets dn2=dt ¼ 0.Thus, a dynamic singlet–doublet equilibrium (s.d.e.) is established:

n2eq ¼�d�Sm

n2eq ð5Þ

At condition (2), it follows from Eq. (5) that

n1eq ffi 1, N1eq ¼ N10 ð6Þ

N2 ¼ ðRevÞN10 or n2 << 1 ð7Þ

Thus, at small values of Rev, the s.d.e. is established with only small devia-tions in the singlet equilibrium concentration from the initial concentration[Eq. (6)]. The doublet concentration is very low compared to the singletconcentration, and the multiplet concentration is very low compared tothe doublet concentration. The last statement follows from a comparison


of the production rates of doublets and triplets. The doublets appear due tosinglet–singlet collisions, whereas the triplets appear due to singlet–doubletcollisions. The latter rate is lower due to the low doublet concentration.The ratio of the number of singlet–doublet collisions to the number ofsinglet–singlet collisions is proportional to Rev.

B. Kinetic Equation for Coupling of Flocculationand Intradoublet Coalescence inMonodisperse Emulsions

Both the rate of doublet disaggregation and the rate of intradoubletcoalescence are proportional to the momentary doublet concentration.This leads (25,31) to a generalization of Eq. (3):

dn2

dt¼

n21�Sm

� n21

�dþ

1

�c

� �

ð8Þ

There are two unknown functions in Eq. (8), so an additional equation isneeded. This equation describes the decrease in the droplet concentrationcaused by coalescence:

d

dtn1 þ 2n2ð Þ ¼ �

n2

�cð9Þ

The initial conditions are

n2jt¼0¼ 0 ð10Þ

dn2

dt

��t¼0

¼n10

�Smð11Þ

Condition (11) follows from Eqs. (8) and (9). The solution of the set ofEqs. (8) and (9) taking into account boundary conditions (10) and (11) isa superposition of two exponents (25,31). In the case

�c >> �d ð12Þ

the solution simplifies (25,31) to

n2ðtÞ ¼�d�Sm

exp �2�dt

�Sm�c

� �

� exp �t

�d

� ��

ð13Þ


Equation (13), as compared to Eqs. (5) and (6), corresponds to the s.d.e. ifthe expression in the second set of brackets equals 1. In the time interval

�d < t < � ð14Þ

where

� ¼ �c�Sm2�d

ð15Þ

the first term in the second set of brackets approximately equals 1, whereasthe second term decreases from 1 to a very small value. Thus, the s.d.e. isestablished during the time �d and preserves during the longer time interval[according to Eq. (14)].

For times longer than �, there is no reason to apply Eq. (13) becausethe condition to linearize Eq. (8) is no longer valid with the concentrationdecrease. At the beginning of the process, the doublet concentrationincreases, whereas later in the process, coalescence predominates andthe doublet concentration decreases. Thus, the function in Eq. (13) has amaximum (25,31).

C. Coalescence in a Singlet–Doublet Systemat Quasiequilibrium

After a time tmax, a slow decrease in the doublet concentration takesplace simultaneously with the more rapid processes of aggregation anddisaggregation. Naturally, an exact singlet–doublet equilibrium is notvalid due to the continuous decrease in the doublet concentration.However, the slower the coalescence, the smaller is the deviation fromthe momentary dynamic equilibrium with respect to the aggregation–disaggregation processes.

It is reasonable to neglect the deviation from the momentarydoublet–singlet equilibrium with the condition

dn2

dt<<

n2

�dð16Þ

Indeed, for this condition, the derivative in Eq. (3) can be omitted, whichcorresponds to s.d.e. characterized by Eq. (5).

It turns out (25,29–31) that the deviation from s.d.e. is negligiblebecause the condition (15) is valid [i.e., for conditions (2) and (12)]. Forthese conditions, the fragmentation of flocs influences the coalescencekinetics, which can be represented as a three-stage process, as illustratedin Fig. 2. During a rather short time �d , the approach to s.d.e. takes place


[i.e., a rather rapid increase in the doublet concentration (stage 1)]. Duringthe next time interval, �d < t < tmax, the same process continues. However,the rate of doublet formation declines due to coalescence (stage 2). The exactequilibrium between the doublet formation and their disappearance due tocoalescence takes place at the time tmax when the doublet concentrationreaches its maximum value, n2ðtmaxÞ. During the third stage, whent > tmax, the rate of doublet fragmentation is lower than the rate offormation, because of the coalescence within doublets. This causes a slowmonotonous decrease in the concentration. Taking into account the s.d.e.[Eqs. (5) and (7)], Eq. (9) can be expressed as

dn1

dt¼ �

�d�Sm�c

n21 ð17Þ

The result of the integration of Eq. (17) can be simplified to

n1ðtÞ ¼n1ðtmaxÞ

1þ n1ðtmaxÞt=2�ffi 1þ

t

2�

� �1

ð18Þ

Figure 2 Three stages in the coupling of aggregation, fragmentation, and

coalescence at the condition �d << �Sm << �c. Initially, the doublet concentration

n2 is very low and the rates of doublet fragmentation and of coalescence are

correspondingly low compared to the rate of aggregation (first stage, no coupling).

Due to increasing n2, the fragmentation rate increases and equals the aggregation

rate at tmax (exact singlet–doublet equilibrium). The growth in n2 stops at tmax

(second stage, coupling of aggregation and fragmentation). Intradoublet coalescence

causes a slight deviation from exact s.d.e. to arise at t > tmax, and the singlet

concentration n1 and the doublet concentration decrease due to intradoublet

coalescence (third stage, coupling of aggregation, fragmentation, and coalescence).

n1 and n2 are dimensionless, (n1 ¼ N1=N10, n2 ¼ N2=N10, N10 is the initial singlet

concentration). (From Ref. 25.)


with a small deviation in n1ðtmaxÞ from 1. Differing from the preceding stageswhen the decrease in the droplet concentration caused by coalescence issmall, a large decrease is now possible during the third stage. Thus, this isthe most important stage of the coalescence kinetics.

D. Reduced Role of Fragmentation with Decreasing �c

With decreasing �c, condition (12) is violated and new qualitative features ofthe destabilization process not discussed in Refs. 29–31 arise. As the ratio�c=�d diminishes and

�c < �d ð19Þ

the s.d.e. is violated because a larger part of the doublets disappears due tocoalescence. Correspondingly, the smaller the ratio �c=�d , the smaller is thefragmentation rate in comparison with the aggregation rate (i.e., the largerthe deviation from s.d.e.). In the extreme case

�c << �d ð20Þ

the fragmentation role in s.d.e. can be neglected. This means that almostany act of aggregation is accompanied by coalescence after a short doubletlifetime. Neglecting this time in comparison with �Sm in agreement withcondition (1), one concludes that any act of aggregation is accompaniedby the disappearance of one singlet:

dn1

dt¼ �

n21�Sm

ð21Þ

This leads to a decrease in the singlet concentration described by anequation similar to the Smoluchovski equation for rapid coagulation:

n1ðtÞ ¼ 1þt

�Sm

� ��1

ð22Þ

The Smoluchowski equation describing the singlet time evolution does notcoincide with Eq. (22). The peculiarity of Eq. (22) is that it describes thekinetics of coupled aggregation and coalescence with a negligible fragmen-tation rate. Due to fragmentation, doublet transformation into multiplets isalmost impossible at condition (1).


The coupling of aggregation, fragmentation, and coalescence in themore general case described by condition (19) leads to

n1ðtÞ ¼n1ðtmaxÞ

1þ n1ðtmaxÞt=2�gffi 1þ

t

2�g

� ��1

t > �d þ �cð Þ ð23Þ

with a small deviation of n1ðtmaxÞ from 1 and

�g ¼�Smð�d þ �cÞ

2�dð24Þ

At conditions (1) and (12) �g � � and Eq. (23) transforms into Eq. (18). Atconditions (1) and (20) �g � �Sm and Eq. (23) transforms into Eq. (22).Equation (24) demonstrates the reduction of the role of fragmentationwith decreasing �c. It is seen that at the transition from condition (19) tocondition (20), �d cancels in Eq. (21) (i.e., the fragmentation rolediminishes).

E. Experimental

1. Video-Enhanced Microscopy (Microslide PreparativeTechnique) for Investigation of Singlet–DoubletEquilibrium and Intradoublet Coalescence (29–31)

Direct observation of doublets in the emulsion bulk is difficult due to doub-lets tending to move away from the focal plane during the time of obser-vation. The microslide preparative technique can, however, be successfullyapplied, providing pseudobulk conditions. A microslide is a plane-parallelglass capillary of rectangular cross section. The bottom and top walls of thecapillary are horizontal, and the gravity-induced formation of a sediment orcream on one of the inner normal surfaces is rapidly completed due to themodest inner diameter of the slide (typically 50–100 mm). If both the volumefraction of droplets in an emulsion and the slide inner diameter are small,the droplet coverage along the inside surface amounts to no more than a fewpercent, and the analysis of results is rather simple. It can be seen throughthe microscope that the droplets which have sedimented (or for the case ofoil droplets, creamed) toward the capillary surface participate in chaoticmotion along the surface. This indicates that a thin layer of water separatingthe surface of the microslide from the droplets is preventing the mainportion of droplets from adhering to the microslide surface—an actionwhich would stop their Brownian motion.


During diffusion along the microslide ceiling, the oil dropletscollide. Some collisions lead to the formation of doublets. Direct visualobservation enables evaluation of the doublet fragmentation time, whichvaries in a broad range (27). Another approach to doublet fragmentationtime determination is based on the evaluation of the average concentrationof singlets and doublets and using the above-outlined theory.

The application of the microslide preparative technique combined withvideo microscopy is promising and has enabled the measurement of thecoupling of reversible flocculation and coalescence (29,31). However,some experimental difficulties were encountered. A modest number ofdroplets could sometimes be seen sticking to the glass surface of themicroslide—an effect that corresponded to electrolyte-induced reductionof electrostatic stabilization by adsorbed surfactant. According to theory,increased salt concentration would increase the number of droplets adheringto the surface, reducing the span of electrolyte concentrations that couldbe used.

2. Improving the Experimental Technique with theUse of Low-Density Contrast Emulsions (30)

The sticking of droplets indicates a droplet–wall attraction and the existenceof a secondary potential pit as that for the droplet–droplet attraction in adoublet. The droplet concentration within the pit is proportional to theconcentration on its boundary. The latter decreases with a decrease in thedensity contrast (i.e., the density difference between the droplet and contin-uous phases). The higher the contrast, the greater the gravity promotedadhesion of droplets to the wall. The electrostatic barrier between the poten-tial pit and the wall retards the rate of sticking. The lower the droplet fluxthrough this barrier, the lower is the potential pit occupancy by droplets.Thus, an essential decrease in the rate of sticking is possible with decreasingdensity contrast.

Oil-in-water emulsions were prepared (30) by mixing dichlorodecane(DCD, volume fraction 1%) into a 5� 10�5 M sodium dodecyl sulfate(SDS) solution with a Silverson homogenizer. The oil phase was a 70:1mixture of DCD, which is characterized by an extremely low-density con-trast to water, and decane.

The droplet distribution along and across the slide was uniform (30).This indicates that there was no gravity-induced rolling either. One slideamong four was examined for 2 weeks without any sticking being observed(30). The absence of the rolling and sticking phenomena allowed acquisitionof rather accurate data concerning the time dependence of the droplet sizedistribution.


3. Measurement of Coalescence Time andDoublet Fragmentation Time

The doublet fragmentation time was measured by direct real-time on-screenobservation of the doublets and by analysis of series of images acquired with1–3-min time intervals (47). The formation and disruption/coalescence of adoublet could thus be determined.

The general form of the concentration dependence agrees with thetheory. At C�3� 10�3 M, both theory and experiment yield times ofabout 1 min; at C¼ 9� 10�3 M, these times exceed 10min. For the calcula-tion of the doublet fragmentation time, the electrokinetic potential wasmeasured (31,48).

In experiments with different droplet concentrations, it wasestablished that the higher the initial droplet concentration, the higher thedoublet concentration. This corresponds to the notion of singlet–doubletequilibrium. However, if the initial droplet concentration exceeds 200–300per observed section of the microslide, multiplets predominate. Both theinitial droplet concentration and size affect the rate of decrease in thedroplet concentration. The larger the droplets, the smaller the concentrationsufficient for the measurement of the rate of decrease in the dropletconcentration. This agrees with the theory of doublet fragmentation time,which increases with droplet dimension. Correspondingly, the probabilityfor coalescence increases. These first series of experiments (29,31) wereaccomplished using toluene-in-water emulsions without the addition ofa surfactant and decane-in-water emulsions stabilized by SDS. The obtaineddata concerning the influence of the electrolyte concentration and surfacecharge density were in agreement with the existing notions about themechanism of coalescence. With increasing SDS concentration and corre-spondingly increasing surface potential, the rate of decrease in the dropletconcentration is reduced.

Two methods were used for the measurement of the coalescence time(30,31). Measurement of the time dependence for the concentrations ofsinglets and doublets and a comparison with Eq. (9) enables an evaluationof the coalescence time. Further, information about the time dependence forsinglets and the doublet fragmentation time may be used as well. Theseresults in combination with Eqs. (18) and (15) determine the coalescencetime. The good agreement between results obtained by these very differentmethods indicates that the exactness of the theory and experiments isnot low.

In recent years, several research groups have improved significantly thetheoretical understanding of coalescence of droplet or bubbles. Thenew results (55–59) together with results of earlier investigations (60–64)


have clarified the role of double-layer interaction in the elementary act ofcoalescence.

Derjaguin-Landau-Verwey-Overbeek (DVLO) theory was applied(65,66) for the description of ‘‘spontaneous’’ and ‘‘forced’’ thinning of theliquid film separating the droplets. These experimental results and DLVOtheory were used (65) for the interpretation of the reported visual study ofcoalescence of oil droplets 70–140 mm in diameter in water over a wide pHinterval. A comparison based on DLVO theory and these experimental dataled the authors to conclude (65) that ‘‘if the total interaction energy is closeto zero or has a positive slope in the critical thickness range, i.e. between 30and 50 nm, the oil drops should be expected to coalesce.’’ In the secondarticle (66), in which both ionic strength and pH effects were studied, coa-lescence was observed at constant pH values of 5.7 and 10.9, when theDebye thickness was less than 5 nm. The main trend in our experimentsand in Refs. 65 and 66 are in accordance, because it was difficult to establishthe decrease in total droplet concentration (t.d.c.) at NaCl concentrationslower than 5� 10�3 M (i.e., DL thicknesses larger than 5 nm). An almostquantitative coincidence in the double-layer influence on coalescence estab-lished in our work for micrometer-sized droplets and in Refs. 65 and 66 foralmost 100 times larger droplets is important for the general knowledgeabout coalescence.

F. Perspective for Generalization of the Theory forCoupling of Coalescence and Flocculation

The proposed theory for coupling of coalescence and flocculation at s.d.e.enables the proposal of some important applications (Sec. VI). At the sametime, generalized theory is necessary, because the role of multiplets increasesafter some time or with a higher initial concentration. At least twoapproaches to this difficult task are seen.

According to our video-microscopic observations, there are largepeculiarities in the structure and behavior of multiplets arising at conditionsclose to s.d.e. These peculiarities can be interpreted as the manifestation ofquasiequilibrium, comprising singlets, doublets, and multiplets. Similar todoublets, the lifetime of triplets, tetraplets, and so forth can be short dueto fragmentation and coalescence. This can be valid for multiplets with an‘‘open’’ structure, in distinction from another structure which can be called‘‘closed.’’ In ‘‘open’’ multiplets, any droplet has no more than one or twocontacts with other droplets, which corresponds to a linear chainlike struc-ture. This causes easy fragmentation, especially for the outermostdroplets within a chain. The ‘‘closed’’ aggregates have a denser and more


isometric structure, in which droplets may have more than two contacts withneighboring droplets. As result, fragmentation is more difficult and thefrequency is lower.

The recent progress in theory of aggregation with fragmentation(67–71) for a suspension creates a premise for a theoretical extensiontoward emulsions. However, the necessity in accounting for coalescencemakes this task a difficult one.

III. COUPLING OF COALESCENCE AND COAGULATION

A. General

For emulsion characterization, the notation n1 represents the numberdensity of single droplets and ni represents the number density of aggregatescomprising i droplets (i¼ 2, 3, . . .). The total number density of singledroplets and all kinds of aggregates is given by

n ¼X1

i¼1

ni ð25Þ

This characterization corresponds to Smoluchowski’s theory (53). To char-acterize coalescence, the total number of individual droplets moving freelyplus the number of droplets included in all kinds of aggregates, nT ,

nT ¼X1

i¼1

ini ð26Þ

is introduced as well.In distinction from Smoluchowski’s theory for suspensions, which

predicts the time dependence of the concentration of all kinds of aggregates,the time dependence for the total droplet number can be predicted at thecurrent state of emulsion dynamics theory.

The quantification of coagulation within the theory of coupledcoagulation and coalescence (CCC theory) is based on the Smoluchowskitheory of perikinetic coagulation. Correspondingly, all restrictions inherentto the Smoluchowski theory of Brownian coagulation are preserved in theCCC theory. This means that creaming and gravitational coagulation aretaken into account. A variant of Smoluchowski’s theory specified withregard for gravitational coagulation is well known (72). However, its appli-cation is very difficult because the rate constant of collisions inducedby gravity depends on droplet dimension (14). Due to the weak particle(aggregate) dimension dependence of the rate constants for Browniancollisions, Smoluchowski’s theory is valid for polydisperse suspensions


and remains valid as polydisperse aggregates arise. Unfortunately, thisadvantage of the Smoluchowski theory can almost disappear when com-bined with the coalescence theory, because the coalescence rate coefficientsare sensitive to droplet dimension. Thus, droplet and aggregate poly-dispersity does not strongly decrease the exactness of the description ofcoagulation in the CCC theory, whereas the exactness of the coalescencedescription can be severely reduced.

Although the coalescence influence on the Brownian coagulation ratecoefficient can be neglected, its influence on the final equations ofthe Smoluchovski theory remains. It can be shown that Smoluchowski’sequation for the total number of particles,

nðtÞ ¼ 1þt

�Sm

� ��1

ð27Þ

remains valid, whereas, in parallel, the equations for the singlet and aggre-gate concentrations cannot be used to account for coalescence. Regardingcoupled coagulation and coalescence, the Smoluchowski equation for n1ðtÞis not exact because it does not take into account the singlet formationcaused by coalescence within doublets.

The coalescence within an aggregate consisting of i droplet is accom-panied by the aggregate transforming into an aggregate consisting of i � 1droplets. Because coalescence changes the aggregate type only, the totalquantity of aggregates and singlets does not change. This means that theSmoluchowski function nðtÞ does not change during coalescence, becauseSmoluchowski defined the total quantity of particles as consisting ofaggregates and singlets.

B. Average Models

Average models do not assign rate constants to each possibility for coales-cence within the aggregates; they deal with certain averaged characteristicsof the process. The models in Refs. 40 and 73 introduce the average numberof drops in an aggregate m, because the number of films in an aggregate nfand m are interconnected. For a linear aggregate,

nf ¼ m� 1 ð28Þ

As the coalescence rate for one film is characterized by ��1c , the decrease in

the average droplet quantity in an aggregate is nf times larger. This is takeninto account in the model of van den Tempel for simultaneous dropletquantity increase due to aggregation and decrease due to coalescence.


Van den Tempel formulates the equation which describes the timedependence for the average number of droplets in an aggregate as

dm

dt¼ KfN10 � �

�1c ðm� 1Þ ð29Þ

where the first term is derived using Smoluchowski’s theory.The total number of droplets nT is the sum of single droplets n1ðtÞ and

the droplets within aggregates,

nT ðtÞ ¼ n1ðtÞ þ nvðtÞmðtÞ ð30Þ

where nv is the aggregate number. The latter can be expressed as

nvðtÞ ¼ nðtÞ � n1ðtÞ ð31Þ

Both terms are expressed by Smoluchowski’s theory. The integration ofEq. (29) and the substitution of the result into Eq. (30) yields the timedependence nT ðtÞ according to the van den Tempel model.

1. The Model of Borwankar et al.

In Ref. 40, the van den Tempel model is criticized and improved throughthe elimination of Eq. (29). The authors point out that the ‘‘incoming’’aggregates which cause the increase in m have themselves undergone coales-cence. This is not taken into account in the first term on the right-hand sideof Eq. (29). Instead of taking a balance on each aggregate (as van denTempel did), Borwankar et al. took an overall balance on all particlesin the emulsion. For linear aggregates, the total number of films in theemulsion is given by

nf nv ¼ ðm� 1Þnv ð32Þ

Thus, instead of Eq. (29), the differential equation for nT follows:

�dnT

dt¼ ��1

c ðm� 1Þnv ð33Þ

where m can be expressed through nT using Eq. (30). The advantage ofthis equation in comparison with Eq. (29) is obvious. However, there isa common disadvantage of both theories, caused by the use of theSmoluchowski equation for n1ðtÞ. Coalescence does not change the total


particle concentration nðtÞ, but it changes n1ðtÞ and, correspondingly, nvðtÞ,according to Eq. (31).

The application of Smoluchowski’s theory in the quantification of thecoupling of coalescence and coagulation has to be restricted with the use ofthe total particle concentration nðtÞ only. The average models of van denTempel and Borwankar et al. do not meet this demand.

The theory of Danov et al. (41) does not contradict this demand,which makes it more correct than the preceding theories. Among theSmoluchowski results, the function nðtÞ only is present in the final equationsof this theory. Although the exactness of averaged models is reduced due tothe violation of the restriction in the use of the Smoluchowski theory, resultsfor some limiting cases are not erroneous.

2. Limiting Cases of Fast and Slow Coalescence

Two limiting cases can be distinguished:

The rate of coalescence is much greater than that of flocculation(rapid coalescence):

��1c >> ��1

Sm ð34Þ

The rate of flocculation is much greater than that of coalescence(slow coalescence):

��1c << ��1

Sm ð35Þ

According to general regularities of physicochemical kinetics, theslowest process is rate controlling. If the coagulation step is rate controlling[viz. when condition (34) is valid], then the coalescence is rapid and thegeneral equation of the theory in Ref. 40 is reduced to second order kinetics[i.e., to Smoluchowski’s equation (27)]. Flocs composed of three, four, andmore droplets cannot be formed because of rapid coalescence within thefloc. In this case, the structure of the flocs becomes irrelevant.

At first glance, the coagulation rate has not manifested itself in theentire destabilization process in the case of slow coalescence [condition (35)].At any given moment, the decrease in the total droplet concentration isproportional to the momentary total droplet concentration (first-orderkinetics), which causes an exponential decrease with time:

n ¼ exp �t

�c

� �

ð36Þ


However, this equation cannot be valid for an initial short period of time,because at the initial moment, there are no aggregates and their quantitycontinues to be low during a short time. This means that the coagulation islimiting during an initial time at any slow coalescence rate. This exampleillustrates the necessity of a more exact approach than that which usesaverage models. This was done by Danov et al. (41).

C. The DIGB Model for SimultaneousCoagulation and Coalescence

This kinetic model proposed by Danov, Ivanov, Gurkov, and Borwankar iscalled the DIGB model here for the sake of brevity. Danov et al. (41)generalized the Smoluchowski scheme (Fig. 3a) to account for dropletcoalescence within flocs. Any aggregate (floc) composed of k particlescan partially coalesce to become an aggregate of i particles (1 < i < k),with the rate constant being Kk,i

c (Fig. 3b). This aggregate is further involvedin the flocculation scheme, which makes the flocculation and coalescenceprocesses interdependent. Therefore, the system exhibiting both flocculationand coalescence is described by a combination of schemes 1 and 2:

dnk

dt¼Xk�1

i¼1

Ki,k�if nink�i � 2

X1

i¼1

Kk,if nkni þ

X1

i¼kþ1

Ki,kc ni �

Xk�1

i¼1

Kk,ic nk

ð37Þ

Figure 3 (a) Model of flocculation according to the Smoluchowski scheme (41).

(b) Coalescence in an aggregate of k particles to become an aggregate of i particles,

with a rate constant Kk,ic , 1 < i < k (41).


Equation (37) is multiplied by k and summed up for all k, which yields theequation for nT , which is expressed through double sums. The change of theoperation sequence in these sums leads to the important and convenientequation

dnT

dt¼X1

k¼1

kX1

i¼kþ1

Ki,kc ni �

X1

i¼2

kXk�1

i¼1

Kk,ic nk ð38Þ

Afterward, a total rate coefficient referring to complete coalescence of theith aggregate

Kic,T ¼

Xi�1

k¼1

ði � kÞKi,kc , i ¼ 2, 3, . . . ð39Þ

is introduced. For linearly built aggregates

Kic,T ¼ K2,1

c ði � 1Þ ð40Þ

is derived. With this expression for Kic,T , using also Eqs. (25) and (26),

Eq. (40) is transformed into

dnT

dt¼ �K2,1

c ðnT � nÞ ð41Þ

The integration result of this first-order linear differential equation is wellknown and is represented in general form without specification of nðtÞ(Eq. (18) in Ref. 41). An interesting peculiarity of this important derivationis the disappearance of terms, related to coagulation at the transition fromthe equation set (37) to the main equation (38). This corresponds to the factthat the total quantity of droplets does not change due to coagulation; itdecreases due to coalescence only.

The coagulation regularity manifests itself in the nðtÞ dependence,arising in Eq. (41). It creates the illusion that Eq. (41) can be specified forany nðtÞ function corresponding to any subprocess affecting the dropletaggregate distribution. For example, the gravitational coagulation theoryleads to a function ngðtÞ (72), but it does not create the opportunity todescribe the gravitational coagulation coupling with coalescence by meansof substituting ngðtÞ into the integral of Eq. (41). Because the coalescenceinfluences the gravitational coagulation, another function has to be substi-tuted into Eq. (41) instead of ngðtÞ. This function has to be derived to


account for the coupling of coalescence and coagulation. One concludes thatEq. (41) cannot be used, because its derivation assumes that the coupling ofgravitational coagulation (or another process) and coalescence is alreadyquantified.

A useful exception is Brownian coagulation and its modeling bySmoluchowski with the coagulation rate coefficients, of which sensitivityto aggregate structure and coalescence is low. The substitution of function(27) into the integral of Eq. (41) yields the equation, characterizing thecoupling of coalescence and Brownian coagulation (41).

In fractal theory (74), it is established that diffusion-limitedaggregates and diffusion-limited cluster–cluster aggregates are built uplinearly. This can simplify the application of the DIGB model. However,the diffusivity of fractal aggregates (75) cannot be described by simpleequations and the Smoluchowski theory. This will cause coagulation ratecoefficient dependence on aggregate structure, decreasing the exactnessof Eq. (41) when applied to fractal aggregates. However, there is noalternative to the DIGB model, which can be used as a crude but usefulapproximation in this case as well. In the absence of an alternative, theDIGB model can be recommended for evaluation in the case of gravita-tional coagulation.

Danov et al. compares their theory with the predictions of averagedmodels for identical conditions. It turns out that if coalescence ismuch faster than flocculation, the predictions of the different models coin-cide. Conversely, for slow coalescence, the results of the averaged modelsdeviate considerably from the exact solution. These two results of the com-parison are in agreement with the qualitative considerations in Section III.B.

Data for the relative change in the total number of droplets as afunction of time are presented in Fig. 4 (41). Figures 4a–4c refer toKFN10 ¼ 0:1 s�1 and the coalescence constant K2,1

c varies between 0:1 s�1

(Fig. 4a) and 0:001 s�1 (Fig. 4c). It is seen that the agreement between theDanov et al. and Borwankar et al. models is better the faster the coalescence,as was explained qualitatively earlier. The van den Tempel curves deviateconsiderably from the other two solutions.

For very long times and irrespective of the values of the kinetic param-eters, the model of Borwankar et al. (40) is close to the numerical solution.This is probably because the longer the time, the smaller the concentrationof single droplets. In this extreme case, the error caused in the averagemodels due to the influence of coalescence on the singlet concentration[not taken into account in the equation for nðtÞ] is negligible.

The shortcomings of the averaged models (40,73) and the advantagesof the DIGB model are demonstrated in Ref. 41. However, the range ofapplicability of this model is restricted by many simplifications and the


Figure 4 Relative change in the total number of droplets versus time: initial number of primary particles

N10 ¼ 1� 1010 cm�3; flocculation rate constant Kf ¼ 1� 10�11 cm3=s; curve 1, the numerical solution

of the set (37); curve 2, the model of Borwankar et al. (40) for diluted emulsions; curve 3, the model of

van den Tempel (73). (a) Coalescence rate constant K2,1c ¼ 1� 10�1 s�1; (b) K2,1

c ¼ 1� 10�2 s�1;

(c) K2,1c ¼ 1� 10�3 s�1. (From Ref. 41.)


neglect of other subprocesses (see Sec. III.A). An efficient analyticalapproach was made possible due to the neglect of the coalescence ratecoefficient’s dependence on the dimensions of both interacting droplets.

The model of Borwankar et al. was examined experimentally inRef. 42. The emulsions were oil-in-water, with soybean oil as the dispersedphase, volume fraction 30%, and number concentration 107–1010 cm�3. Theemulsions were gently stirred to prevent creaming during the aging study.A sample was placed on a glass slide, all aggregates were broken up, and thesize of the individual droplets was measured. A rather good agreement withthe theory was established. However, the fitting of the experimental datawas accomplished using two model parameters, namely the coalescence andcoagulation rate coefficients. For the last coefficient, the optimal values(different for two emulsions) were obtained, strongly exceeding theSmoluchowski theory value (Sec. II.A). An interpretation is that ortho-kinetic and perikinetic coagulation took place simultaneously due to stir-ring. Several experiments are known (discussed in Ref. 56) whichdemonstrate better agreement with the value for the coagulation constantpredicted in the Smoluchowski theory.

IV. DOUBLET FRAGMENTATION TIME

A. Theory of Doublet Fragmentation Time

A doublet fragmentation was described by Chandrasekhar (76) by thediffusion of its droplets from the potential minimum, characterizing theirattraction. The time scale for this process takes the form (77)

�d ¼6��a3

�Texp

�Umin

�T

� �

ð42Þ

where Umin is the depth of the potential minimum.To derive the formula for the average lifetime of doublets, Muller (78)

considered the equilibrium in a system of doublets and singlets; that is, thenumber of doublets decomposing and forming are equal. Both processes aredescribed by the standard diffusion flux J of particles in the force field of theparticle that is regarded as central.

Each doublet is represented as an immovable particle with the secondsinglet ‘‘spread’’ around the central one over a spherical layer, whichcorresponds to the region of the potential well. The diffusion flux J of‘‘escaping’’ particles is described by equations used in the Fuchs theoryof slow coagulation. The first boundary condition corresponds to theassumption that the escaping particles do not interact with other singlets.


The second condition reflects the fact that the potential well contains exactlyone particle.

At a small separation between the droplets in a doublet, thedroplet diffusivity reduces because of the increasing hydrodynamicresistance during the droplet approach. A convenient interpolation formulawas used (78) for the description of the influence of hydrodynamic interac-tion on the mutual diffusivity. The difference between the more exactMuller equation and Eq. (42) is caused mainly because of this hydrodynamicinteraction.

B. Doublet Fragmentation Time of Uncharged Droplets

In this subsection, we consider a doublet consisting of droplets with anonionic adsorption layer. The closest separation between two dropletssurfaces h0 exceeds the double thickness of the adsorption layer (2ha). Asa crude approximation, h0 can be identified with 2ha. In the case of smallsurfactant molecules, 2ha� 2 nm.

In this case, the potential well has a sharp and deep minimum. Thismeans that the vicinity of this minimum determines the value of the integral(42). For examination of this assumption, Eq. (42) was calculated numeri-cally and according to the approximate equation (28)

Z�

�

’ðtÞ exp½ f ðtÞ� dt ¼2�

f 00ðtmÞ

� �1=2’ðtmÞ exp½ f ðtmÞ� ð43Þ

where tm corresponds to the potential well minimum.The difference in results was small and enabled application of Eq. (43)

for the calculation and substitution of the asymptotic expression (13,16):

UðhÞ ¼ �A

12

a

hð44Þ

which is valid at small distances to the surface. The result of calculationsaccording to Eqs. (43) and (44) (the Hamaker constant A ¼ 1:3� 10�20 J )are shown in Fig. 5. The chosen value of the Hamaker constant is consistentwith those reported elsewhere (79,80). In addition to the value ofA ¼ 1:3� 10�20 J, we mention other values of the Hamaker constantemployed elsewhere. For example, in food emulsions (80), the Hamaker


constant lies within the range of 3�10�21 J to 10�20 J. The results of calcula-tions for smaller Hamaker constants are also presented in Fig. 5.

The influence of the adsorption layer thickness on doublet lifetime isshown in Fig. 6 for one value of the Hamaker constant. There is highspecificity in the thickness of a polymer adsorption layer. �-Caseinadsorbed onto polystyrene latex causes an increase in the radius of theparticle of 10–15 nm (81). A layer of �-lactoglobulin appears to be in theorder of 1–2 nm thick, as compared to 10 nm for the caseins (82).

When adsorbed layers of hydrophilic nature are present, the repulsivehydration forces must be taken into account. At low ionic strengths, therepulsion follows the expected exponential form for double-layer interaction:

UðhÞ ¼ Kse�ðh=hsÞ ð45Þ

In Ref. 83, the authors emphasize that the surface charge in food emulsions islow, electrolyte concentrations are high, and, hence, the DL is not responsiblefor emulsion stability. The stabilization can be caused by the hydrationforces. However, the flocculation to the secondary minimum remains.Meanwhile, this conclusion must be specified with account for dropletdimension.

Figure 5 Dependence of doublet lifetime on droplet dimension at different values

of the Hamaker constant A: curve 1: A ¼ A1 ¼ 1:33� 10�20 J; curve 2: A ¼ 0:5A1;

curve 3: A ¼ 0:35A1; curve 4: A ¼ 0:25A1; curve 5: A ¼ 0:1A1. The shortest

interdroplet distance is 2 nm. (From Ref. 28.)


C. Lifetime of a Doublet of Charged Droplets andCoagulation/Flocculation

As seen in Fig. 1 of Ref. 39, the coordinates of the secondary minimumcorresponds to �hmin ¼ 5–12 nm. Due to this rather large distance, the fre-quency dependence of the Hamaker constant may be of importance, and theHamaker function AðhÞ characterizing molecular interaction should beintroduced.

In Ref. 84, the distance-independent interaction at zero frequency andinteraction at nonzero frequency is considered separately:

AðhÞ ¼ A0 þ ½AðhÞ � A0� ð46Þ

The result from 36 systems in Ref. 84 are in a rather good accordance with thecalculations of other papers. According to Churaev, the system polystyrene–water–polystyrene can be used to estimate the Hamaker function for oil–water systems. However, with increasing droplet separation, the importanceof A0 is increasing because of AðhÞ � A0. The component A0 is screened inelectrolyte concentrations, because of dielectric dispersion (85–87). At a dis-tance of �hmin � 3�5 nm, the authors (86) found that molecular interaction

Figure 6 The adsorption layer thickness influence on the droplet lifetime of

an uncharged droplet. Adsorption layer thickness: curve 1: h0 ¼ 1 nm; curve

2: h0 ¼ 2 nm; curve 3: h0 ¼ 4 nm; curve 4: h0 ¼ 6 nm. A ¼ 1:33� 10�20 J. (From

Ref. 28.)


disappeared at zero frequency. Experimental evidence concerning this state-ment is discussed in Ref. 16. When evaluating the secondary minimum coa-gulation, A0 can be omitted, as illustrated in Ref. 87.

For illustration of the influence of electrolyte concentration,Stern potential, and particle dimension, some calculations of doublet lifetimeare made and their results are presented in Fig. 7. The potential welldepth increases and, in parallel, doublet lifetime increases with increasingparticle dimension and electrolyte concentration and decreasing surfacepotential.

V. COALESCENCE COUPLED WITH EITHERCOAGULATION OR FLOCCULATION INDILUTE EMULSIONS

Limited attention is paid to the role of fragmentation in emulsion science.A comparison of the prediction of coalescence with and without accountingfor fragmentation (Secs. II and III) enables evaluation of the significance offragmentation. This comparison will be done in Section V.A.

The theories of Refs. 41 and 25 have different areas of applicability(not specified in the articles) and are complementary. Naturally, this

Figure 7 The dependence of doublet lifetime on the Stern potential for different

electrolyte concentrations and droplet dimensions. Numbers near the curves

correspond to droplet radius. Curves 1–4 without account for retardation of

molecular forces of attraction, � ¼ e =kT . Curves 10– 40 with account for

retardation. (From Ref. 28.)


complicates the choice between these theories with respect to concrete con-ditions of the experiments. An approximate evaluation of the aforesaidareas of applicability is given in Section V.B.

A. Fragmentation of Primary Flocs in Emulsions and theSubsequent Reduction of Coalescence

Floc fragmentation reduces the quantity of interdroplet films and, corre-spondingly, retards the entire coalescence process. This retardation can becharacterized by the comparison of Eq. (18) with the theory of Ref. 41,which neglects fragmentation. The longer the time, the greater the retarda-tion, which enables the use of the simpler theory of Ref. 40 for comparison.The results for longer times coincide with the predictions of the more exacttheory of Ref. 41.

The results of the theory of Ref. 41 concerning slow coalescence areillustrated by curve 1 in Fig. 3c in Ref. 41, which is redrawn here as Fig. 8a.It can be seen that for a low value of the coalescence rate constant,the semilogarithmic plot is linear, indicating that the process follows acoalescence rate-controlled mechanism according to Eq. (36). Differingfrom the simple exponential time dependence in Eq. (36), second-orderkinetics dominate at rapid doublet fragmentation, even if coalescence isvery slow. The physical reason becomes clear when considering howEq. (18) is derived. As seen from Eq. (17), the rate of decline in the dropletconcentration is proportional to the doublet concentration. The latter isproportional to the square of the singlet concentration at s.d.e., whichcauses second-order kinetics. Thus, at slow coalescence, the disaggregationdrastically changes the kinetic law of the coalescence (i.e., from the expo-nential law to second-order kinetics).

In the second stage, coagulation becomes the rate-controlling processbecause of the decrease in the collision rate accompanying the decrease inthe droplet concentration. Thus, at sufficiently long times, second-orderkinetics characterizes both reversible and irreversible aggregation.Nevertheless, a large difference exists even when identical functions describethe time dependence, as the characteristic times are expressed throughdifferent equations for irreversible and reversible aggregation. In the firstcase, it is the Smoluchowski time; in the second case, it is the combination ofthree characteristic times [i.e., Eq. (18)].

Let us now try to quantitatively characterize the reduction in coales-cence caused by doublet disintegration. For this purpose, the calculationsare performed according to Eq. (6) at �Sm ¼ 10 s and �c ¼ 103 s (Fig. 8a),102 s (Fig. 8b), and 10 s (Fig. 8c). For all figures, the same value of the ratio


2�d=�Sm ¼ 0:1 is accepted, satisfying condition (2). In all of these figures, thecalculations according to Eq. (18) are illustrated by curve 2.

The comparison of curves 1 and 2 characterizes the reduction of coales-cence caused by doublet disintegration; the lower the Rev values, the strongerthe reduction. The simple curve 1 in Fig. 8a can be used also for higher �cvalues, because then the condition (12) is even better satisfied. Thus, if �c1 andt1 correspond to the data of Fig. 8a and �c2 ¼ m�c1 with m >> 1, the identity

�c2t2 ¼ �c1mt1

mð47Þ

Figure 8 Relative change in the total number of droplets versus time; initial

number of droplets N10 ¼ 1� 1010 cm�3; flocculation rate constant

Kf ¼ 1� 10�11 cm3=s; curve 1—calculations according to Eq. (18); curve 2—the

model of Borwankar et al. (40) for dilute emulsions coalescence rate constants:

(a) K2,1c ¼ 1� 10�1 s�1; (b) K2,1

c ¼ 1� 10�2 s�1; (c) K2,1c ¼ 1� 10�3 s�1. Coalescence

times: (a) �c ¼ 103 s; (b) �c ¼ 102 s; (c) �c ¼ 10 s. Smoluchowski time �Sm ¼ 10 s.

Doublet lifetime �d ¼ 0:5 s. nT is the dimensionless total droplet concentration,

nT ¼ NT=N10. (From Ref. 25.)


is useful. This means that

nT

n0�c1m,

t1

m

� ¼

nT

n0ð�c1, t1Þ ð48Þ

[i.e., t2 ¼ t1=m, where the right-hand side of Eq. (48) is drawn in Fig. 9]. Forexample, Fig. 9 is similar to Fig. 8a and can be used for 100 fold longer time,shown on the abscissa axis. The increase in �c enables us to increase �Smwithout violating condition (35) and with Eq. (36) valid. Thus, �Sm ¼ 1000 sor lower can be chosen as condition for Fig. 9. Curve 2, characterizing therate of doublet disintegration, preserves as well if the value of 2�d=�Sm ¼ 0:1remains; now, it corresponds to a higher �d value of 5 s.

B. Domains of Coalescence Coupled Either withCoagulation or with Flocculation

The condition

Rev >> 1 ð49Þ

corresponds to coagulation. A theory for the intermediate case

Rev � 1 ð50Þ

when part of the droplets participate in flocculation and another coagulateis absent. To specify the conditions (2) and (49), the doublet lifetime must be

Figure 9 Similar to Fig. 8, with other values for the characteristic times.

Coalescence time �c ¼ 105 s; the Smoluchowski time �Sm ¼ 103 s; Doublet fragmen-

tation lifetime �d ¼ 50 s. (From Ref. 25.)


expressed through surface force characteristics (viz. through the surfaceelectric potential and the Hamaker function) and droplet dimension, aswas described in Section IV.

In the equation for the Smoluchowski time [Eq. (4)], the dropletnumerical concentration N10 can easily be expressed through the dropletvolume fraction ’ and the average droplet radius a (we replace a polydis-perse emulsion by an ‘‘equivalent’’ monodisperse emulsion). The resultinganalysis respective to a and ’ is easier than relating to N10 because theboundary of application of different regularities are usually formulatedrespective to a and ’. The Smoluchowski time is

�Sm ¼ ��1F ’

�1 4

3�a3 ð51Þ

We exclude from consideration a special case of extremely dilute emulsions.Comparing Fig. 7 and the results of calculations according to Eq. (51),one concludes that condition (49) is mainly satisfied. It can beviolated if simultaneously the droplet volume fraction and the dropletdimension are very small. This occurs if ’ < 10�2 and a < 0:2�0:3 mm.Discussing this case, we exclude from consideration the situation whena 0:1 mm, corresponding to microemulsions and ’ << 10�2. With thisexception, one concludes that for uncharged droplets, flocculationis almost impossible because condition (2) cannot be satisfied. A secondconclusion is that at

a < 0:2�0:3 mm ð52Þ

the theory in Ref. 41 cannot be applied without some corrections madenecessary by the partially reversible character of the aggregation. Themain conclusion is that when

a > 0:2 mm and ’ > 10�2 ð53Þ

the theory in Ref. 41 does not need corrections respective to the reversibilityof flocculation. However, this conclusion will change at the transition toa thicker adsorption layer. As described in Section IV, the thicker theadsorption layer, the shorter is the doublet fragmentation time.

The electrostatic repulsion decreases the depth of the potential welland, correspondingly, decreases the doublet lifetime. As result, flocculationbecomes possible for submicrometer droplets as well as for micrometer-sizeddroplets if the electrolyte concentration is not too high, the surface potential


is rather high, and the droplet volume fraction is not too high. This is seenfrom Fig. 7.

The reversibility criterion depends on many parameters in the case ofcharged droplets. To discriminate and to quantify the conditions ofcoagulation and flocculation, let us consider Rev values lower than 0.3 aslow and values higher than 3 as high. In other words, coagulation takesplace when Rev>3, whereas at Rev<0.3, there is flocculation; that is, theconditions

Rev ¼�dðc0, ’, aÞ

�Smða, ’Þ> 3 ð54Þ

Rev < 0:3 ð55Þ

determine the boundaries for the domains of coagulation and flocculation.These domains are characterized by Fig. 10 and correspond to fixed valuesof the droplet volume fraction. In addition, a definite and rather largedroplet dimension 2a ¼ 4 mm is fixed. After fixation of the values of

Figure 10 Domains of coagulation and flocculation. Curves 1 and 2 are calculated

with the Rabinovich–Churaev Hamaker function; a twice higher value is used for

the calculation of curves 10 and 20. The domain of flocculation is located above

curve 1, while the domain of coagulation is located beneath curve 2. Volume

fractions: ’ ¼ 0:01 (Fig. 10a) and ’ ¼ 0:1 (Fig. 10b); particle dimension 2a ¼ 4mm:(From Ref. 28.)


volume fraction and droplet dimension, the domains are characterized incoordinates � and C.

In Fig. 10, the domain of flocculation is located above and to the leftof curve 2; the domain of coagulation is located beneath and to the right ofcurve 1. To indicate the sensitivity of the domain boundaries to theHamaker function value, curves 10 and 20 are calculated using values twotimes higher than those of curves 1 and 2.

In distinction from uncharged droplets, flocculation in the range ofmicrometer-sized droplets is possible. As seen in Fig. 10, even rather largedroplets (4 mm) aggregate reversibly if the electrolyte concentration is lowerthan ð1�5Þ � 10�2 M and the Stern potential is higher than 25mV. Forsmaller droplets, the domain of flocculation will extend while the domainof coagulation shrinks. For submicrometer droplets, flocculation takes placeeven at high electrolyte concentrations (0.1M).

C. Hydration Forces Initiate Flocculation

Due to the similar dependence on the distance h of hydration forcesand electrostatic interaction, the decrease of doublet lifetime caused byhydration forces of repulsion can be calculated because of this similarity.It is sufficient to use the substitution hs for �

�1 and Ks for

16"kT

e

� �2

tanhe

kT

� �2

ð56Þ

where k is the Boltzmann constant, T is absolute temperature, and e is theelementary charge. The doublet lifetime can be determined with the use ofthe results presented in Fig. 7. For the sake of brevity, a similar figure withKs given on the ordinate axis and hs on the abscissa axis is not shown.It turns out that the decrease in �d caused by hydration forces leads toflocculation of submicrometer droplets. As for micrometer-sized droplets,coagulation takes place with the exception for the case when both hs and Ks

are rather large.

VI. APPLICATIONS

The restrictions in Eqs. (1) and (15) corresponding to strong retardation ofthe rate of multiplet formation and slow intradoublet coalescence are notfrequently satisfied. Nevertheless, these conditions are important becausethey correspond to the case of very stable emulsions. Because the kinetics ofretarded destabilization of rather stable emulsions are of interest, attention


has to be paid to provide these conditions and, thus, the problem of coupledcoalescence and flocculation arises.

There are large qualitative distinctions in the destabilization processesfor the coupling of coalescence and coagulation, and coalescence and floc-culation. In the first case, rapid aggregation causes rapid creaming andfurther coalescence within aggregates. In the second case, the creaming ishampered due to the low concentration of multiplets, and coalescence takesplace both before and after creaming. Before creaming, singlets predominatefor a rather long period of gradual growth of droplet dimensions due tocoalescence within doublets. The discrimination of conditions for couplingof coalescence with either flocculation or coagulation is accomplished inRef. 28.

The creaming time is much shorter in the case of coagulation and,correspondingly, the equation describing the coupling of coalescence andflocculation preserves its physical sense for a longer time than is the casefor coagulation. One concludes that the theory of the coupling of coalescenceand flocculation provides a new opportunity for the long-term predictionof emulsion stability, although creaming restricts the application ofthis theory as well. Note that this restriction weakens in emulsions of low-density contrast and in water-in-oil (w/o) emulsions with a high-viscositycontinuum.

Long-term prediction is a two-step procedure. The first step is thedetermination of whether an emulsion exhibits coagulation or flocculation.It means that the characteristic time �d must be measured and comparedwith �Sm, the value of which is easily evaluated taking into account themeasured concentration using Eq. (4). A comparison of these times enablesthe choice between condition (1) and the opposite condition (�Sm << �d).The second step is the prediction of the evolution in time for the t.d.c. Ifcondition (1) is valid, Eq. (18) has to be used for the prediction [� in Eq. (18)has to be specified in accordance with Eq. (15)]. In the opposite case, DIGBtheory must be used.

A. Long-Term Prediction of Emulsion Stability

It is possible, in principle, to give a long-term prediction of emulsion stabil-ity based on the first indications of aggregation and coalescence. The nextexample clarifies the principal difficulty in a reliable long-term prediction ifa dynamic model of the emulsion is not available.

The first signs of aggregation and coalescence can always becharacterized by a linear dependence if the investigation time t is shortcompared with a characteristic time � for the evolution of the total droplet


concentration nðtÞ:

nðtÞ ¼ n0 1�t

�

� ð57Þ

This short-time asymptotic corresponds to many functions [e.g., to Eq. (18)or to Eq. (36)]. The first can arise in the case of coalescence coupled withcoagulation (41), whereas the second can arise for coalescence coupled withflocculation (31). The discrimination between irreversible and reversibleaggregation is only one component of emulsion dynamics modeling(EDM) and it is seen that without this discrimination, the difference inthe prediction of the time necessary for a droplet concentration decrease(e.g., 1000 times) can be 7� and 1000�.

B. Refinement of Methods for Emulsion Stabilization(Destabilization) by Means of the Effect on BothCoalescence and Flocculation

Emulsion stability (or, for that matter, instability) can be described from theviewpoint of the coupling of coalescence and flocculation. However, foremulsifiers (or demulsifiers), only their influence on the elementary act ofcoalescence is primarily taken into account. The coupling of coalescence andflocculation is reflected in Eq. (15) and one concludes that it follows themultiplicativity rule and not the additivity rule. This means that the totalresult of the application of a stabilizer (destabilizer) depends very much onboth flocculation and fragmentation. The development of a more efficienttechnology for emulsion stabilization (destabilization) is possible by takinginto account the joint effect on both the coalescence and the aggregation(disaggregation) processes.

1. Combining Surfactants and Polymers inEmulsion Stabilization

The coalescence rate depends mainly on the thin-film (black) stability andcorrespondingly on the short-range forces, whereas flocculation depends onthe long-range surface forces. Due to this important difference, synergism inthe dependence of these processes on the different factors can be absent. Theuse of a single surfactant only may not provide, at the same time, both theoptimal fragmentation and optimal stability of an emulsion film. Probablythe use of a binary surfactant mixture with one component which providesthe film stability and a second one which prevents the flocculation mayprovide optimal emulsion stabilization. Naturally, coadsorption of the


two is necessary. For such an investigation, a measurement method for boththe doublet fragmentation time and the coalescence time is necessary.

2. There Is a Strong Influence of Low Concentrations ofIonic Surfactant on Doublet Fragmentation Timeand Coalescence Time

Let us consider the situation when an emulsion is stabilized against coales-cence by means of an adsorption layer of nonionic surfactant and is stronglycoagulated because of the subcritical value of the Stern potential that isusual for inorganic electrolytes (48) at moderate pH. In a large floc, anydroplet has many neighbors, meaning a rather high number of interdropletfilms per droplet. The coalescence rate is proportional to the total number offilms and can be rather high. It can be strongly decreased by adding evena low concentration of an ionic surfactant. This can be sufficient to providea supercritical Stern potential value that will be accompanied by a drasticdecrease in the doublet lifetime compared to that of weakly chargeddroplets.

At shorter doublet lifetimes, flocculation can become reversible and itcan stop at the stage of singlet–doublet equilibrium. It will provide a strongdecrease in the coalescence rate because coalescence occurs within doubletsonly and their concentration can be very low.

Thus, a modest addition of an ionic surfactant to an amount of anonionic surfactant sufficient to provide an almost saturated adsorptionlayer can make the overall emulsion stabilization more efficient.The nonionic surfactant suppresses coalescence but cannot preventflocculation, whereas the ionic surfactant retards the development offlocculation.

We can give an example when both coalescence and flocculation areaffected by an ionic surfactant (SDS). In Ref. 88, it is established thatcoalescence is suppressed at SDS concentrations exceeding 6� 10�5 M:Meanwhile, the CCC is 2� 10�2 M NaCl at 10�6 M SDS. Thus, SDSconcentrations slightly above 10�6 M are sufficient to retard flocculation.In this example, it is essential that the concentrations needed toretard flocculation are very low compared to those needed to preventcoalescence.

It is noteworthy that low concentrations of an ionic surfactantcan increase emulsion stability due to the simultaneous manifestation ofthree mechanisms. First, the depth of the secondary potential minimumdecreases due to the electrostatic repulsion that is accompanied by a�d decrease. Second, the transition from the secondary minimum throughan electrostatic barrier and into the primary minimum extends the


coalescence time. Third, the time of true coalescence (i.e., the timenecessary for thin-film rupture) increases due to electrostatic repulsionas well (29,65).

C. Standardization of the Measurement of �c and �d

Direct investigation of the coalescence subprocess in emulsions is difficult.Instead, the entire destabilization process is usually investigated.Meanwhile, the rate of the destabilization process depends on the rates ofboth flocculation and disaggregation and on the floc structure as well. Allthese characteristics vary in a broad range. Given an unknown value for thetime of the elementary act of coalescence, �c, the different times can bemeasured for the integrated process and different evaluations of �c arethen possible.

The rate of coalescence in an aggregate essentially depends on thenumber of droplets within it and the packing type (i.e., on the number offilms between the droplets). This complication is absent when consideringthe case of the s.d.e.

The possible advantage of �c measurement at s.d.e. is in avoiding thedifficulty caused by polydispersity of droplets appearing during precedingcoalescence within large flocs. At s.d.e., the initial stage of the entirecoalescence process can be investigated when the narrow size distributionof an emulsion is preserved.

At s.d.e., determination of the time dependence of the t.d.c. issufficient for the investigation of coalescence. In Refs. 29 and 30, this wasaccomplished through direct visual observation. By using video-enhancedmicroscopy and computerized image analysis, the determination of the t.d.c.can be automated. Such automated determination of total droplet numberin a dilute DCD-in-water emulsion at the s.d.e. can be recommended asa standard method for the characterization of the elementary act ofcoalescence.

In parallel, the second important characteristic (viz. the doubletfragmentation time) is determined by the substitution of �c, �Sm, andmeasured �d into Eq. (18).

D. Experimental–Theoretical Approach toEmulsion Dynamics Modeling

1. General

To predict the evolution of the droplet (floc) size distribution is the centralproblem concerning emulsion stability. It is possible, in principle, to predict


the time dependence of the distribution of droplets (flocs) if informationregarding the main subprocesses (flocculation, floc fragmentation, coales-cence, creaming), constituting the whole phenomenon, is available. This pre-diction is based on consideration of the population balance equation (PBE).

The PBE concept was proposed by Smoluchowski. He specified thisconcept for suspensions and did not take into account the possibility of flocfragmentation. Even with this restriction, he succeeded in the analyticalsolution neglecting gravitational coagulation and creaming, and he obtainedthe analytical time dependence for a number of aggregates ni comprisingi particles (i¼ 2, 3, . . .).

In the most general case, the equation for the evolution of the totaldroplet number takes into account the role of aggregation, fragmentation,creaming, and coalescence. There is no attempt to propose an algorithmeven for a numerical solution to such a problem.

The usual approach in the modeling of an extremely complicatedprocess is the consideration of some extreme cases with further synthesisof the obtained results. The next three main simplifications are inherent tothe current state of emulsion dynamics modeling: the neglection of theinfluence of the gravitational field (i.e., neglection of creaming/sedimenta-tion); in a first approximation, it is possible to consider either coagulation orflocculation; finally, neglection of the rate constant dependence on dropletdimension.

2. Combined Approach in Investigations of Dilute andConcentrated Emulsions

The modeling of collective processes in concentrated emulsions is ex-tremely complicated. Recently, the efficiency of computer simulation in thesystematic study of aggregates, gels, and creams has been demonstrated (50).Monte Carlo and Brownian dynamics are particularly suited to thesimulation of concentrated emulsions. However, information aboutdroplet–droplet interaction is necessary. The reliability of this informa-tion is very important to provide reasonable results concerning concen-trated emulsions. In other words, the assumption concerning pairwiseadditive potentials for droplet–droplet interaction and the thin-emulsion-film stability must be experimentally confirmed. The extraction of thisinformation from experiments with concentrated emulsions is very diffi-cult. On the other hand, measurement of the doublet fragmentation timein dilute emulsions is a convenient method to obtain information aboutpairwise additive potentials. Information about pair potentials and theelementary act of coalescence obtained in experiments with diluteemulsions preserves its significance for concentrated emulsions as well.


One concludes that modeling of concentrated emulsions becomespossible by combining experimental investigation of the simplest emulsionmodel system with computer simulation accounting for the characteristics ofa concentrated emulsion (high droplet volume fraction, etc.).

3. In the Transformation of PBE into an Efficient Tool for EDM,the Determination of Kernels Is the Main Task

The levels of knowledge concerning kernels describing different subpro-cesses differ strongly. There exists a possibility for quantification of kernelsrelated to aggregation and fragmentation (14,39,78). On the other hand, thecurrent state of knowledge is not sufficient for prediction of the thin-filmbreakdown time.

The deficit in knowledge about thin-film stability currently makespurely theoretical modeling of emulsion dynamics impossible. As a result,a complex semitheoretical approach to EDM is necessary. The PBE is themain component of both the experimental and the theoretical stages of thisapproach. In the experimental stage, the PBE simplified for s.d.e. providesthe background for the determination of the coalescence kernels with the useof experimental data (29,30).

For the determination of the coalescence kernels, the more compli-cated reverse task must be solved—namely their determination based onthe comparison of experimental data about the emulsion evolution in timewith solution of the PBE. In the absence of an analytical solution, thereverse task is usually very difficult. The most efficient way to overcomethis difficulty is an experimental realization with the use of the universallysimplest conditions for emulsion time evolution, which can be describedanalytically.

4. The Simplest Emulsion State for Which InvestigationCan Provide Information About Coalescence Is atSinglet–Doublet Quasiequilibrium with SlowCoalescence Within the Doublets

The simplest singlet–doublet emulsion can exist at singlet–doublet quasi-equilibrium given slow coalescence within doublets. Its simplicity resultsin a very simple kinetic law for the entire kinetics of coupled flocculationand coalescence, [viz. Eq. (18)]. Thus, s.d.e. provides the most convenientconditions for investigations of the elementary act of coalescence and thedoublet fragmentation time.

The main simplification in all existing models for emulsion dynamics(25,41) is the neglection of the coalescence time dependence on droplet


dimensions. This simplification is not justified and reduces severely the valueof the prediction, which can now be made with use of the PBE. For elim-ination of this unjustified simplification, it is necessary to determine thecoalescence time for emulsion films between droplets of different dimensionsi and j (viz. �cij, similar to the existing analytical expressions for the doubletfragmentation time, �dij) (14). The determination of a large set of �cij valuesby means of a comparison of experimental data obtained for an emulsionconsisting of different multiplets and the PBE numerical solutions for it isimpossible. On the other hand, this paramount experimental–theoreticaltask can be solved for a dilute emulsion at s.d.e. and slow intradoubletcoalescence.

5. Substitution of the Coalescence Kernels Makes thePBE Equation Definite and Ready for the Predictionof Emulsion Time Evolution (With the Restriction ofLow-Density Contrast and Without Accounting forGravitational Creaming and Coagulation)

With application of the scaling procedure for the representation of thekinetic rate constants for creaming and gravitational coagulation, the PBEis solved analytically in Ref. 89. This scaling theory creates a perspective forthe incorporation of creaming in the emulsion dynamics model in parallelwith coalescence, aggregation, and fragmentation.

VII. SUMMARY

The mechanisms of kinetic stability in macroemulsions and miniemulsionsare completely different. Strong droplet deformation and flattening of theinterface in a macroemulsion cause the Reynolds mode of drainage, whichprolongs the life of the emulsion. This mechanism is not importantfor miniemulsion droplet interaction, because either the deformation andflattening are weak (charged droplets) or the Reynolds drainage is rapid dueto the small dimension of the interdroplet film (uncharged droplets). Kineticstability in a miniemulsion can result from floc fragmentation if the electro-kinetic potential is not too low and the electrolyte concentration is not toohigh, corresponding to a degree of electrostatic repulsion.

The potential strength of physicochemical kinetics with respectto emulsions is the PBE, enabling prediction of the time evolution of thedroplet size distribution (d.s.d.) when the subprocesses [including dropletaggregation, aggregate fragmentation, droplet coalescence and droplet (floc)creaming] are quantified. The subprocesses are characterized in the PBE by


the kinetic coefficients. The coupling of the four subprocesses, the dropletpolydispersity, and the immense variety of droplet aggregate configurationscauses extreme difficulty in EDM. The three processes of aggregation,fragmentation, and creaming can be quantified. In contrast, only theexperimental approach is now available for efficient accumulation ofinformation concerning emulsion film stability and coalescence kernelquantification.

Correspondingly, EDM may be accomplished by combiningexperiment and theory: (a) the determination of coalescence and fragmenta-tion kernels with the use of emulsion stability experiments at low-densitycontrast (l.d.c.) and s.d.e., because this allows for the omittance of creamingand gravitational terms in PBE, simplifying the equation and enabling kerneldetermination; (b) the prediction of the droplet size evolution as functionof time by means of solution of the PBE, specified for the determinedcoalescence and fragmentation kernels. This mathematical model has to bebased on the PBE supplemented by terms accounting for the role of creamingand gravitational coagulation in the aggregation kinetics.

Emulsion dynamics modeling with experiments using l.d.c. emulsionsand s.d.e. may result in the following: (a) The quantification of emulsionfilm stability [viz. the establishment of the coalescence time dependenceon the physicochemical specificity of the adsorption layer of a specificsurfactant (polymer), its structure, and the droplet dimensions]. Such quan-tification can form a basis for optimized selection and synthesis of emulsi-fiers and demulsifiers for a broad variety of technological applicationsof emulsions, replacing the current empirical approach dominating thisarea. (b) The elaboration of a commercial device for coalescence timemeasurement, in combination with EDM, will represent a useful approachto the optimization of emulsion technology with respect to stabilization anddestabilization.

ACKNOWLEDGMENT

The technology program FLUCHA, financed by industry and theNorwegian Research Council, is acknowledged for financial support.

REFERENCES

1. D. G. Dalgleish, in Emulsions and Emulsion Stability (J. Sjoblom, ed.),

Marcel Dekker, New York, 1996, pp. 287–325.

2. D. G. Dalgleish, in Encyclopedic Handbook of Emulsion Technology (J. Sjoblom,

ed.), Marcel Dekker, New York, 2001, pp. 207–232.


3. P. Becher (ed.), Encyclopedia of Emulsion Technology, Volume 1, Marcel

Dekker, New York, 1983.

4. P. Becher (ed.), Encyclopedia of Emulsion Technology, Volume 2, Marcel


5. J. Sjoblom (ed.), Emulsions—A Fundamental and Practical Approach, Kluwer

Academic, Dordrecht, 1992, p. 1.

6. R. A. Mohammed, A. I. Bailey, P. F. Luckham, and S. E. Taylor, Colloids

Surfaces 80, 223 (1993); 80, 237 (1993); 83, 261 (1994).

7. C. I. Chitewelu, V. Hornof, G. H. Neale, and A. E. George, Can. J. Chem. Eng.

72, 534 (1994).

8. L. Bertero, A. DiLullo, A. Lentini, and L. Terzi, in Proceedings of the SPE

Annual Technical Conference and Exhibition (Part 2) (SPE 28543) Society of

Petroleum Engineers, New Orleans, 1994, p. 283–295.

9. J. F. McCafferty and G. G. McClaflin, in Proceedings of the SPE Annual

Technical Conference, SPE 24850, Society of Petroleum Engineers, 1992,

pp. 857–868.

10. J. A. Kitchener and P. R. Musselwhite, in Emulsion Science (I. P. Sherman, ed.),

Academic Press, New York, 1968.

11. J. Sjoblom (ed.), Emulsions and Emulsion Stability, Marcel Dekker, New York,

1996.

12. I. B. Ivanov and P. A. Kralchevsky, Colloids Surfaces 128, 155 (1997).

13. B. V. Derjaguin, Theory of Stability of Colloids and Thin Films, Plenum Press,

New York, 1989.

14. S. S. Dukhin and J. Sjoblom, in Emulsions and Emulsion Stability (J. Sjoblom,

ed.), Marcel Dekker, New York, 1996, p. 41.

15. I. Gregory, Crit. Rev. Environ. Control. 19, 185 (1989).

16. J. N. Israelachvili, Intermolecular and Surface Forces, Academic Press, London,

1991.

17. J. Lyklema, Fundamentals of Interface and Colloid Science, Academic Press,

London, 1993, Vol. 1.

18. P. A. Kralchevsky, K. D. Danov, and N. D. Denkov, in Handbook of Surface

and Colloid Chemistry (K. S. Birdi, ed.), CRC Press, London, 1996.

19. Reference deleted.

20. R. K. Prud’homme (ed.), Foams: Theory, Measurement and Applications,

Marcel Dekker, New York, 1995.

21. P.A.Kralchevsky,K.D.Danov, and I. B. Ivanov, inFoams: Theory,Measurement

and Applications (R. K. Prud’homme, ed.), Marcel Dekker, New York, 1995.

22. S. A. K. Jeelany and S. Hartland, J. Colloid Interf. Sci. 164, 296 (1994).

23. P. D. I. Fletcher, in Drops and Bubbles in Interface Research (D. Mobius and

R. Miller, eds.), Elsevier, New York, 1998, Vol. 6.

24. P. J. Breen,D.T.Wasan,Y.H.Kim,A.D.Nikolov, andC. S. Shetty, inEmulsions

and Emulsion Stability (J. Sjoblom, ed.), Marcel Dekker, New York, 1996.

25. S. S. Dukhin and J. Sjoblom, J. Dispers. Sci. Technol. 19, 311 (1998).

26. Ø. Holt, Ø. Sæther, J. Sjoblom, S. S. Dukhin, and N. A. Mishchuk, Colloids

Surfaces 123–124, 195 (1997).


27. Ø. Sæther, S. S. Dukhin, J. Sjoblom, and Ø. Holt, Colloid J. 57, 836 (1995).

28. S. S. Dukhin, J. Sjoblom, D. T. Wasan, and Ø. Sæther, Colloids Surfaces 180,

223 (2001).

29. Ø. Sæther, J. Sjoblom, S. V. Verbich, N. A. Mishchuk, and S. S. Dukhin,

Colloids Surfaces 142, 189 (1998).

30. Ø. Sæther, J. Sjoblom, S. V. Verbich, and S. S. Dukhin, J. Dispers. Sci. Technol.

20, 295 (1999).

31. Ø. Holt, Ø. Sæther, J. Sjoblom, S. S. Dukhin, and N. A. Mishchuk, Colloids

Surfaces 141, 269 (1998).

32. S. S. Dukhin, Ø. Sæther, J. Sjoblom, Proc. Special Symposium in honor of D.T.

Wasan on Emulsions, Foams, Thin Films and Interfacial Rheology as part of 72nd

ACS Colloid and Surface Science Conference, 1998.

33. O. Reynolds, Phil. Trans. R. Soc. (London) A 177, 157 (1886).

34. A. Scheludko, Adv. Colloid. Interface Sci. 1, 391 (1967).

35. K. D. Danov, D. N. Petsev, N. D. Denkov, and R. Borwankar, J. Chem. Phys.

99, 7179 (1993).

36. K. D. Danov, N. D. Denkov, D. N. Petsev, and R. Borwankar, Langmuir 9,

1731 (1993).

37. D. N. Petsev, N. D. Denkov, and P. A. Kralchevsky, J. Colloid Interf. Sci. 176,

201 (1995).

38. I. B. Ivanov and D. S. Dimitrov, in Thin Liquid Films (I. B. Ivanov, ed.),

Marcel Dekker, New York, 1998.

39. N. A. Mishchuk, J. Sjoblom, and S. S. Dukhin, Colloid J. 57, 785 (1995).

40. R. P. Borwankar, L. A. Lobo, andD. T.Wasan,Colloids Surfaces 69, 135 (1992).

41. K. D. Danov, I. B. Ivanov, T. D. Gurkov, and R. P. Borwankar, J. Colloid

Interf. Sci. 167, 8 (1994).

42. L. A. Lobo, D. T. Wasan, and M. Ivanova, in Surfactants in Solution (K. L.

Mittal and D. O. Shah, eds.), Plenum Press, New York, 1992, Vol. 11, p. 395.

43. L. Lobo, I. Ivanov, and D. Wasan, Mater. Interf. Electrochem. Phenomena 39,

322 (1993).

44. G. Narsimhan and E. Ruckenstein, in Foams: Theory, Measurement and

Applications (R. K. Prud’homme, ed.), Marcel Dekker, New York, 1995.

45. D. Exerova, D. Kashchiev, and D. Platikanov, Adv. Colloid Interf. Sci. 40,

201 (1992).

46. B. V. Derjaguin and Yu. V. Gutop, Kolloid. Zh. 24, 431 (1962).

47. A. V. Prokhorov and B. V. Derjaguin, J. Colloid Interf. Sci. 125, 111 (1988).

48. S. V . Verbich, S. S. Dukhin, A. Tarovsky, Ø. Holt, Ø. Sæther, and J. Sjoblom,

Colloids Surfaces 23, 209 (1997).

49. D. Kashchiev and D. Exerowa, J. Colloid Interf. Sci. 203, 146 (1998).

50. E. Dickinson and S. R. Euston, Adv. Colloid Interf. Sci. 42, 89 (1992).

51. P. Taylor, Adv. Colloid Interf. Sci. 75, 107 (1998).

52. H. W. Yarranton and Y. H. Masliyah, J. Colloid Interf. Sci. 196, 157 (1997).

53. M. Smoluchowski, Phys. Z. 17, 557 (1916).

54. G. A. Martynov and V. M. Muller, Kolloidn. Zh. 36, 687 (1974).

55. V. M. Muller, Kolloidn. Zh. 40, 885 (1978).


56. H. Sonntag and K. Strenge, Coagulation Kinetics and Structure Formation, VEB

Deutcher Verlag der Wissenschaften, Berlin, 1987.

57. J. D. Chan, P. S. Hahn, and J. C. Slattery, AIChE J. 34, 140 (1988).

58. S. A. K. Jeelani and S. Hartland, J. Colloid Interf. Sci. 156, 467 (1993).

59. R. W. Aul and W. I. Olbriht, J. Colloid Interf. Sci. 115, 478 (1991).

60. A. Scheludko and D. Exerova, Colloid J. 168, 24 (1960).

61. A. Scheludko, Proc. K. Ned. Akad. Wet. Ser. B 65, 87 (1962).

62. D. Platikanov and E. Manev, in Proceedings, 4th International Congress of

Surface Active Substances, Plenum, New York, 1964, p. 1189.

63. K. A. Burrill and D. R. Woods, J. Colloid Interf. Sci. 42, 15 (1973).

64. K. A. Burrill and D. R. Woods, J. Colloid Interf. Sci. 42, 35 (1973).

65. S. R. Deshiikan and K. D. Papadopoulos, J. Colloid Interf. Sci. 174, 302 (1995).

66. S. R. Deshiikan and K. D. Papadopoulos, J. Colloid Interf. Sci. 174, 313 (1995).

67. M. E. Costas, M. Moreau, and L. Vicente, J. Phys. A: Math. Gen.

28, 2981 (1995).

68. F. Le Berre, G. Chauveteau, and E. Pefferkorn, J. Colloid Interf. Sci. 199,

1 (1998).

69. B. J. McCoy and G. Madras, J. Colloid Interf. Sci. 201, 200 (1998).

70. J. Widmaier and E. Pefferkorn, J. Colloid Interf. Sci. 203, 402 (1998).

71. I. M. Elminyawi, S. Gangopadhyay, and C. M. Sorensen, J. Colloid Interf. Sci.

144, 315 (1991).

72. M. von Smoluchowski, Phys. Chem. 92, 129 (1917).

73. M. van den Tempel, Rec. Trav. Chim. 72, 419 (1953); 72, 433 (1953).

74. P. Meakin, Adv. Colloid Interf. Sci. 28, 249 (1988).

75. R. Nakamura, Y. Kitada, and T. Mukai, Planet. Space Sci. 42, 721 (1994).

76. S. Chandrasekhar, Rev. Mod. Phys. 15, 1 (1943).

77. W. B. Russel, D. A. Saville, and W. R. Schowalter, Colloidal Dispersions,

Cambridge University Press, New York, 1989.

78. V. M. Muller, Colloid J. 58, 598 (1996).

79. Ia. I. Rabinovich and A. A. Baran, Colloids Surfaces 59, 47 (1991).

80. P. Walstra, in Gums and Stabilizers for the Food Industry (G. O. Philips, P. A.

Williams, and D. J. Wedlock, eds.), IRL, Oxford, 1988, Vol. 4, p. 233.

81. D. G. Dalgleish, Colloids Surfaces 46, 14 (1990).

82. D. G. Dalgleish and Y. Fang, J. Colloid Interf. Sci. 156, 329 (1993).

83. B. Bergenstahl and P. M. Claesson, in Food Emulsions (K. Larson and

S. Friberg, eds.), Marcel Dekker, New York, 1989, p. 41.

84. Ya. I. Rabinovich and N. V. Churaev, Kolloidn. Zh. 46, 69 (1984); 52,

309 (1990).

85. D. J. Mitchell and P. Richmond, J. Colloid Interf. Sci. 46, 128 (1974).

86. V. N. Gorelkin and V. P. Smilga, Kolloidn. Zh. 34, 685 (1972).

87. V. N. Gorelkin and V. P. Smilga, in Poverkhnostnye Sily v Tonkikh Plenkakh i

Ustoichivost Kolloidov (B. V. Derjaguin, ed.), Nauka, Moscow, 1974, p. 206.

88. S. Usui, Y. Imamura, and E. Barough, J. Dispers. Sci. Technol. 8, 359 (1987).

89. S. B. Grant, C. Poor, and S. Relle, Colloids Surfaces 107, 155 (1996).


6Structure and Stability of AeratedFood Emulsions

D. T. Wasan, W. Xu, A. Dutta, and A. NikolovIllinois Institute of Technology, Chicago, Illinois, U.S.A.

I. INTRODUCTION

The development of a stable structure and a suitable product texture in foodemulsions and foam-based food products depends on the interactionsbetween the fat globules and between fat globules and air bubbles (1). Animportant factor influencing the stability of food emulsions and foams isthe developed fat particle structure. The fat particle structure variationand destabilization during the shearing process are very complex. Manyfactors such as shear rate, shear time, temperature, the composition of thesurfactant mixture, and the ratio of liquid to solid content inside a fatglobule can greatly affect the fat particle structure of the final product.Researchers have investigated some of these factors in relationship toemulsions.

Long-term stability of whipped foams is thought to result from theformation of a three-dimensional fat particle network structure (2,3). Thepresence of solid fat (to promote fat globule rupture) and of liquid fat (topromote clumping) might be necessary for the formation of a stablewhipped foam. Darling (4) obtained similar results. He found that a pre-requisite for effective whipping was that the solid part of the fat and thewhipping time were directly related to the proportion of solid fat. Brookerand his colleagues (5,6) found that the development of a stable structure in awhipped foam depends on the interactions between the fat globules andbetween fat globes and air bubbles. The stabilization of air was a two-stepprocess involving the selective adsorption of proteins and emulsions to theair-bubble surface, followed by the attachment of fat globules to the


air–water interface under shear stress. The attached globules stayed at theair–water interface in such a way that part of the fat became exposed tothe air and slightly protruded into it. In the final product, bubbles werestabilized predominantly by the fat globules, but the remains of the initialprotein-adsorbed interface persisted between the globules. In the aqueousphase, the fat particles form a three-dimensional network structure connect-ing different bubbles, which decreases the mobility of the bubbles and,hence, mechanically stabilizes the whipped product.

However, most information about the stability of whipped foams orfoam-based products has come from studies utilizing simplified solutionsthat neglected the effect of the dispersed phase or from empirical studiesproviding information that was not easily translated quantitatively. The lackof detailed information from actual whipped foams stems from the lack ofmethods or tools to measure the component functional properties in systemsas complex as whipped foams, which has prevented a further understandingof the relationship between component interactions (structure and confor-mation) and the stability of whipped foams. We have successfully appliedthe backlight-scattering technique to quantitatively measure the fat particlestructure inside the food systems in a noninvasive way (7).

It has been generally accepted that interfacial properties such as sur-face and interfacial tension, diffusion, and adsorption–desorption propertiesof surfactants and proteins are of paramount importance to the stability offoam-based food products (4,8–13). Researchers have expended much effortin determining the static and dynamic interfacial properties, and a numberof methods exist to study that. However, all of these methods are based onthe concept that the foam or emulsion stability is controlled by the inter-facial properties of a single interface in isolation. However, several of theinterfaces are in close proximity to the surfaces in these aerated emulsionsystems, with a film of only a few nanometers to almost a few micrometersin thickness. Under these conditions, there are interactions occurring acrossthe film that can be significant. Toward this end, we have recently developeda new noninvasive method to measure the static and dynamic properties ofthe foam and emulsion films (14–18).

Dispersed air may degrade in many ways; for example, it may degradethrough aggregation, creaming, coalescence, or Ostwald ripening and gasdiffusion (19). Ostwald ripening is the process by which larger bubbles growat the expense of the smaller ones because of differences in their chemicalpotential. The growth occurs by the diffusion of the dispersed phase throughthe continuous phase. Liftshitz and Slezov (20) and, independently, Wagner(21) developed an analytical solution for the Ostwald ripening process for adilute dispersed phase. Gas diffuses out of the system in an open systemwhere the dispersed phase is gas, and it results in a decrease in the bubble


size and overrun. De Vries (22,23) and Princen et al. (24,25) have providedthe basic framework for studying gas diffusion to the atmosphere. Prins (26)showed the effect of the viscoelasticity of the bubble surface on the rate ofgas diffusion. Recent studies (27–29) have been done to understand theeffect of bulk rheology on bubble dissolution. However, very little workhas been done to understand the effects of gas diffusion and Ostwaldripening on the stability of aerated food emulsions.

The purpose of this chapter is to provide quantitative descriptions ofthe fat particle structure variation during the whipping process using thebacklight-scattering technique and to relate this structure to the stability ofthe whipped products. This research also attempts to interpret the dynamicfilm tension, film elasticity, and film thickness stability of foam-based pro-ducts and to understand the effects of Ostwald ripening and gas-diffusionmechanisms on the long-term stability of aerated food products.

II. MATERIALS AND METHODS

A. Emulsion Preparation and Whipping

The emulsion preparation is as follows: Sodium caseinate, emulsifiers (poly-sorbate 60 and sorbitan monostearate), and gums (xanthan and guar) weredispersed in water. Partially hydrogenated vegetable oil, syrups, and flavorswere then added to make a preemulsion, which was then pasteurized andhomogenized in a two-stage homogenizer. The resulting emulsion wascooled to 1–3�C and aged for 45min in a tank at about 5�C to producethe final emulsion samples used in this study. Air was injected into the agedemulsion under pressure, and the mixture was whipped under high stress for15min. The final whipped foam was produced after equilibrium was reachedinside contherm 1 and contherm 2 (Fig. 1). Two food emulsion samples wereprepared by this method. The fat concentrations in these two samples are12wt% and 20wt%, respectively.

The sizes of the fat particles inside the food emulsion samples weremeasured using a Horiba LA-900 particle size distribution analyzer (HoribaInstruments Incorporated, Irvine, CA). The fat particle sizes in these twofood emulsion samples were nearly the same (mean particle size around0.4 mm).

B. Mesophase Preparation

The aerated mesophase was chosen as a model system to study therole of gas diffusion and Ostwald ripening on the stability of anaerated food system. The mesophase consisted of 2wt% surfactant,


33 wt% sugar, and 65wt% water. We chose propylene glycol stearate[hydrophile–lipophile balance (HLB): 1.8; molecular weight (MW): 387]and triglycerol stearate (HLB: 7.2; MW: 568) as low-HLB surfactantsand sucrose stearate (HLB: 15; MW: 652) as a high-HLB surfactant. Alow-HLB surfactant is dry mixed with sucrose stearate in the desiredproportions (i.e., a high-HLB to low-HLB surfactant molar ratio waskept as 0.56) and added, under vigorous stirring, to water maintained ata temperature of 80�C. The total surfactant concentration was kept at6wt%. The surfactants were mixed for about 15min in order to ensurethe complete dissolution of the surfactants in the solution. The meso-phase was refrigerated overnight prior to use. At the time of the experi-ment, the mesophase was mixed with equal weights of sugar and water.The mesophase was whipped using a Hobart (N-50) whipping machine.Whipping was done for 1min at low speed followed by a high speedof aeration for 2min at room temperature (22–25�C). The aeratedmesophase was stored at room temperature and in refrigerated conditions.We measured the bubble size distribution and overrun as a function oftime in order to describe the destabilization process quantitatively.

Figure 1 Sketch of typical food emulsion and foam process.


C. Backlight-Scattering Experiment

The nondestructive backlight-scattering technique has been successfullyapplied to obtain the information of the particle structure and interparticleinteractions inside colloidal dispersions in our previous research (7,30). Inthis research, the backlight-scattering technique was used to obtain theinformation about the particle structure variation under and after theshear stress. Figure 2 illustrates the experimental setup: a collimated greenlight beam (50 mm in diameter, wavelength 543 nm) produced by a 0.5-mWHe–Ne laser (Model 155A, Spectra-Physics Laser Products, MountainView, CA) was scattered by colloidal particles inside the sample under acontrolled temperature and produced a scattering image on the surface of aglass vial. The image intensity was then recorded by a vertically polarizedcoupled-charge device (CCD) digital camera (Lynxx 2000 Frame 336*244FT,Spectra Source Instruments, Westlake Village, CA) and downloaded to a

Figure 2 Backlight-scattering experimental setup.


computer (IBM-compatible PC Pentium/120MHZ). The recorded imagewas transformed into an intensity profile by image-analyzing software(Lynxx 2000, Version 3.04b, Spectra Source Instrument, Westlake, CA).According to our previous research (7), the value of the normalized struc-ture factor S(Q) [which is defined as the structure factor S(Q)max at the firstmaximum point divided by the structure factor S(Q)min at the first minimumpoint] can be used to quantify the fat particle structure inside food emul-sions due to the decreased oscillatory behavior of the structure factoragainst the scattering vector. Larger values of S(Q) yield more orderedstructures inside colloidal dispersions.

D. Film Rheology and Film Thickness Stability

When the bubbles or droplets interact in a foam or emulsion system, a filmis formed from the continuous phase between the bubbles or drops. Thestability of any foam or emulsion depends on the response of the thin liquidfilm and the Plateau borders during shear and dilatation. In real polydis-persed foam and emulsion systems, thin liquid films formed between bub-bles or drops are not flat, but have a spherical curved shape. We developed adynamic film tensiometer (Fig. 3) to study the film’s rheological propertiesand thickness stability (14–18, 30,31). A small drop of emulsion (15–20 mmin size) is placed at the tip of a glass capillary (0.6mm of inner diameter and0.05mm of wall thickness) and a drop with a double meniscus is formed.The capillary is connected to a feed syringe with a piston having a finescrew. A curved, spherical, cap-shaped film is formed with its meniscusadhering to the capillary drop by expelling the emulsion using air (Fig. 3).The diameter of the film for the foam lamella is slightly larger than thecapillary diameter. Because the inner and the outer parts of the foam areair, the surface of the foam lamella is a part of a sphere. A sensitive pressuretransducer (with sensitivity of 0.25 dyn/cm2) was used to measure thecapillary pressure versus time. The output of the pressure transducer isfed into a computer using a data acquisition card. The size of the film iscontrolled (i.e., expanded or contracted) by the feed syringe.

The interfacial film tension ( f ) is related by the Young–Laplaceequation to the film radius (Rf) and capillary pressure (Pc) as

Pc ¼2f

Rfð1Þ

The film tension is equal to the two times the single interfacial tensionbecause the film consists of two single interfaces.


Dynamic experiments are usually conducted by either expanding orcontracting the film area. In the film stress–relaxation experiments, the filmis quickly expanded or contracted, and then the relaxation of the film ten-sion (capillary pressure) is measured. During this process, the film area iskept constant. When the film area expands, the surfactant concentration onthe film surfaces drops. The film tension and the capillary pressure jumpfrom the equilibrium value of the surface tension to the value correspondingto the new area per surfactant molecule on the film surfaces. The surfactantsin the meniscus and in the film start to adsorb to restore equilibrium and,thus, the capillary pressure and film tension decreases as time progresses andfinally approach an equilibrium value. The initial film tension ( fi ) versus thelogarithm of the relative film expansion area provides information aboutGibbs film elasticity as follows:

Ef ¼dfi

d lnðA=A0Þð2Þ

where A0 is the initial film area.

Figure 3 Diagram of the film rheometer.


Our film rheometric technique also allows for the study of the dynamicfilm thickness stability of curved films. Due to the decreased surfactantconcentration on the film expansion surface, critical film expansion ratioexists, and after that, the films will rupture.

E. Foam Film Air Permeability

The stability of foam-based food products depends on the air permeabilitythroughout the foam lamella (32). In order to measure the foam film airpermeability, a curved, spherical, cap-shaped foam film was formed at thetip of a glass capillary (0.6mm inner diameter and 0.05mm wall thickness),and the foam bubble diameter and capillary pressure were monitored versustime. The temperature was kept constant at 5�C during the measurement.Foam film permeability can be calculated by using the air flow rate, capil-lary pressure, foam lamella area, viscosity, and foam lamella thickness. Thefoam film tension measurement by the film rheometer provides simulta-neous information about the capillary pressure, gas flow rate, and foamlamella area as a function of time. The gas flow rate (q) throughout thefoam lamella is related by the Darcy equation to the film area (A),the capillary pressure (Pc), viscosity m, foam lamella thickness (h), andpermeability (k) (33):

q ¼ kAPc

�h

� �

ð3Þ

When the foam lamella is formed on the tip of capillary, the lamella thins.During that time, the air diffuses throughout the foam lamella and the foamcell shrinks. The foam lamella increases its thickness due to the viscosityeffect during the shrinking. The rate of the foam cell shrinking process ismodeled assuming that the film thinning rate (Vf) follows Reynolds law:

Vf ¼2h3F

3��r41ð4Þ

where r1 is the film radius of the curved film, h is the film thickness, and F isthe capillary force. As the film drains, the thickness decreases from theinitial thickness h1 to thickness h2. The time that the film thickness decreasesfrom h1 to h2 is given by the Reynolds equation:

t12 ¼3��r414F

1

h22�

1

h21

� �

ð5Þ


Assuming that the total film volume is constant during foam cell shrinking(i.e., high film viscosity as in food emulsions samples), then

h2 ¼R1L1h1

R2L2ð6Þ

where R1 is the foam cell radius before shrinking, R2 is the foam film radiusafter shrinking, L1 is the foam cell height before shrinking, and L2 is thefoam cell height after shrinking.

Combining Eqs. (5) and (6), the film thickness of the shrinking foamfilm as a function of time can be derived as

h2 ¼ h2,drainage þ h2,shrinkage � h1 ð7Þ

where

h2,shrinkage ¼R1L1h1

R2L2ð8Þ

and

h2,drainage ¼1

h21þ4�t12PcAf

3��r41

� ��1=2

ð9Þ

F. Measurement of Overrun and Bubble Size

Overrun of the mesophase is defined by the percentage of incorporated airwith respect to the initial mesophase volume. Incorporated air reduces thedensity of the mesophase. The amount of incorporated air can be estimatedby measuring the change in the density of the mesophase. Measurementerror was estimated to be about 5–10% in this method. However, this isan absolute error (i.e., if the overrun is 1000%, then the confidence intervalis 990–1010%). Bubble size distributions were measured from the videoframes using a transmitted light microscope and VCR, and diameter mea-surements were taken using ImagePro Software (images from the videotapecan be captured by this software). These captured images can be analyzedautomatically or manually. Bubbles are clearly seen in this image. We usedthe manual mode to measure the diameter of the bubbles; one can accuratelymeasure bubbles having a diameter of 1 mm or more. (The presence ofsmaller bubbles can be observed visually.) Aerated samples were sandwichedbetween a glass slide and a thin glass plate. The thin glass plate was verylight and probably did not affect the bubble shape as only a very smallnumber of distorted bubbles were observed. This is probably the result of


the high elasticity of the mesophase, which could sustain the weight of thethin glass. As an additional measure, we measured the diameters of bubblesthat appear distorted and took the average. We then focused on a singleplane of the mesophase. The bubbles from the other planes are out of focusand cannot be seen very clearly. We did not measure the bubbles that wereout of focus. A substantial number of bubbles (100–150) were counted ineach frame. In order to obtain statistical estimates of bubble size distribu-tion, 6–10 frames (containing a total of about 800–900 bubbles) were used.

G. Measurement of Gas Diffusion

The rate of air loss from a single bubble is measured to quantify the gas loss(decrease in overrun) from the system. A bubble is formed inside the liquid(i.e., mesophase) and the diameter of the bubble is measured as a function oftime. This would give an idea of the permeability of the surfactant film to airand permeability of air through the liquid medium. The pressure differencebetween the bubble and the atmosphere is the main cause for gas transferfrom the bubble to the atmosphere. The rate of gas loss depends on the sizeof bubble, the surface tension of the liquid, the permeability of the liquidfilm, and the distance from the free gas–liquid surface. Air bubbles are eitherat the gas–liquid interface or in the bulk liquid. In order to simulate anaerated mesophase, we decided to study two cases: (I) gas diffusion froma bubble when the bubble is at the gas–liquid interface and (II) gas diffusionfrom a bubble when the bubble is inside the bulk liquid. The bubble shrinksover time when the bubble is formed at the gas–liquid interface. Let usimagine that a bubble of radius r is formed at the gas–liquid interface.The pressure difference between the bubble and the atmosphere will be

�P ¼2

rð10Þ

where is the surface tension.Now, the concentration difference can be expressed in terms of the

pressure difference. If we combine the above equation with the ideal gas lawand Fick’s law of diffusion, the following equation (24) can be obtained:

r2 ¼ r20 � 4�D

hP0

� �

t ð11Þ

where r is the radius at the instant time t, r0 is the initial radius, P0 is theatmospheric pressure, D is the diffusivity of the film, h is the thickness of thefilm, is the surface tension, and a is the dimensionless factor (24). � is ameasure of the area of the bubble that is exposed to the air directly above


the plane of contact with the liquid. This equation relates the radius of ashrinking bubble with time. Almost all of the gas diffuses through the filmarea when the bubble is at the gas–liquid interface (directly exposed to theatmosphere). The rate of gas diffusion depends on the area of the film, thethickness of the film, and the film structure (molecular packing in the film).

The physics of the problem remains the same when the bubble is insidethe bulk liquid. Let us imagine that a bubble of radius r is placed inside theliquid at a distance h from the gas–liquid interface. If we neglect the pressureexerted by the liquid height h, the pressure difference between the bubbleand the atmosphere can be expressed by Eq. (10). The concentration differ-ence can be expressed in terms of the pressure difference using Henry’s law.By combining Eqs. (10) and (11) with the ideal gas law and Henry’s law, weobtain

r2 ¼ r20 � 4RT

P0

� �DS

h

� �

t ð12Þ

Equation (12) relates the radius of a shrinking bubble with time when thebubble is inside the liquid; S is the solubility of the dispersed phase (air) atthe interface.

A bubble is formed inside the mesophase using the setup describedin Fig. 4 in order to measure the gas diffusion from a bubble inside themesophase. The bubble is kept at a fixed distance from the air–mesophaseinterface using a thin glass plate. The distance of the bubble from theair–mesophase interface was approximately 250 mm. The diameter of thebubble is measured as a function of time to determine the rate of air lossfrom a single bubble.

III. RESULTS AND DISCUSSION

A. Fat Particle Structure Variation During Homogenization

Homogenization is responsible for the formation of food emulsions. Thestructure factor was measured immediately after homogenization for threeemulsions containing 5wt%, 12wt%, and 20wt% fat. Figure 5 showsthat the structure factor for all the samples was approximately 1, indi-cating that a very poor fat-particle-packing structure was produced afterhomogenization; therefore, all of the food emulsions were unstable.

The temperature is very high after homogenization (around 72.5�C).Xu et al. (7) observed that a food emulsion become less stable with anincrease in temperature. Figure 5 also reveals that the fat-particle-orderingstructure was better (more than 1) for an emulsion containing 20wt% fat.


The fat particle concentration inside a 20-wt% fat emulsion was muchhigher than that of other two samples and the increased fat particle con-centration probably resulted in improved fat particle structure.

B. Fat Particle Structure During Cooling

The emulsions were cooled to nearly 5�C to prevent fat particle aggregationor coalescence after homogenization. The fat particle structuring factorgreatly improved after cooling the three samples when compared with theparticle size distribution after homogenization. It was found that theparticle size distribution remained unchanged in all three samples. The nor-malized structure factor for an emulsion containing 20wt% fat is shown inFig. 6. Increased fat particle structuring suggests that the food emulsionsshould be cooled down immediately after homogenization to avoid fat par-ticle aggregation or coalescence. The particle size distribution for all threesamples remained unchanged. Starch particles were added after coolingand before aging. The fat particle structure becomes less ordered due tothe addition of large starch particles for the 5-wt% fat whipped foam.The effect of starch particles and fat destabilization has been systematically

Figure 4 Experimental setup for the single-bubble experiment.


Figure 6 Normalized structure factor after cooling for an emulsion containing

20wt% fat.

Figure 5 Normalized structure factor after homogenization.


studied by Xu et al. (7). It was found that the effect of starch on the fatparticles was similar to that of particle polydispersity. Increased particlepolydispersity decreased the particle ordering (7), and the stability of thefood emulsion decreases.

The fat particle structure for the 12-wt% fat and 20-wt% fat whippedfoams are slightly greater than the other sample and benefited from theadsorption of surfactants and proteins on the fat particle structure.However, the increase in the fat particle structuring was small, implyingthat most of the adsorption of surfactants and protein on the fat particlesurface was completed before aging.

C. Fat Particle Structure During Aging

It was observed that the fat particle structure for emulsions containing5-wt% and 12-wt% fat did not change with aging time. For an emulsioncontaining 20wt% fat, a small and gradual increase was observed (Fig. 7).No significant change was observed in the fat particle size distributionduring the aging process. Some studies (4,34) have concluded that one

Figure 7 Fat particle structure variation during aging for an emulsion containing

20wt% fat.


important role of aging was to provide the necessary time to achieve thecomplete coverage of fat globules with adsorbed molecules and competitiveadsorption between surfactant and proteins. However, our results suggestthat most of the adsorption of surfactant and proteins was completed beforethe aging process. Based on our observations and the results of the otherstudies (4,34), the role of aging can be summarized as follows:

1. The polysaccharide stabilizers, such as gums, require time forfull hydration; therefore, the bulk viscosity increases with theaging time.

2. The crystallization of fat particles occurs during the aging process.According to Darling and Goff (34), the existence of crystalline fatinside fat particles was important to the whippability of foodemulsions. They found that without the crystalline phase, theundesired coalescence of fat droplets occurs during the whippingprocess, and the resulting whipped foams are unstable.

D. Fat Particle Structure During Aeration

It was observed that the fat particle structure inside the low-fat emulsions(5wt% and 12wt%) did not change during the aeration process. However,the fat particle structure inside the high-fat product (20% fat) decreased.This is probably the result of the increased collision between the fat particlesat high fat concentrations due to high mechanical agitation. Increased fatparticle collisions resulted in flocculation and clustering, which, in turn,reduced the fat-particle-ordering structure.

E. Fat Particle Structure During the Whipping Process

It was observed that the fat particle structuring was greatly improved(Figs. 8–10) during the whipping process, indicating a well-developed fatparticle structuring in the bulk phases. It is worth noting that the normalizedstructure factor after whipping in our samples is found to be the same as thesequence of stability of whipped foam products (i.e., 20wt% fat whippedfoam >5wt% fat whipped foam >12wt% fat whipped foam), indicatingthat many factors, such as fat particle attachment at the air–water interfaceand the formation of a three-dimensional fat particle network structurebetween neighboring air bubbles, can influence the stability of whippedfoams. The developed fat particle structure in the bulk phase after whippingis one of the most important factors in determining the stability of whippedfoams.

The data shown in Figs. 8–10 also indicate that in a stable whippedfoam, most of the fat particles in the bulk phase remain in a single particle


Figure 8 Fat particle structure variation during whipping process (for an emulsion

containing 5wt% fat).

Figure 9 Fat particle structure variation during whipping process (for an emulsion

containing 12wt% fat).


state and are well distributed in space, rather than being flocculated,clustered, clumped, or even coalesced to form large particles. The effect offat particle floccules, clusters, clumps, or coalesced fat particles on the fatparticle structure is similar to the addition of particle polydispersity. Whenfat particle aggregation occurs, the fat particle structure should become lessordered (7). Other researchers reached a similar conclusion using varioustechniques (2,3,5,6,11). They found that the outline of the individual fatglobules and clusters in a stable whipped foam were clearly observable; inthe unstable dairy foam, the fusion/coalescence of fat particles wasobserved.

Controlled fat particle destabilization is necessary in order to producestable whipped foams. For example, before fat particles can attach to theair–water interface or the three-dimensional fat particle network structurecan form between neighboring air bubbles, the fat particles have to beruptured or destabilized (clustered or clumps). The fat particle destabiliza-tion during the whipping process should be optimally controlled to producea stable whipped foam. The fat particle structure decreased with the increas-ing fat particle concentration in the second whipping stage (contherm 2). Infact, the fat particle structure decreased for the 20-wt% fat whipped foam,indicating that too many fat particles were destabilized and overwhipping

Figure 10 Fat particle structure variation during whipping process (for an

emulsion containing 20wt% fat).


occurred. In this case, only one whipping stage is necessary for high-fatproducts.

The stability of the whipped foams depends on the interactionsbetween the fat globules and air bubbles. This leads to the buildup of amatrix in which air bubbles are stabilized. The initial stage of the whippingprocess involves the adsorption of protein and surfactants onto the air–water interface. This gives an initial stabilizing effect on the foams.Subsequent phases of the whipping process involve a progressive increasein the extent to which the fat particles are whipped to attach at the air–waterinterface, the formation of a fat particle structure in the bulk phase, thebuildup of a three-dimensional network structure between neighboring airbubbles, and changes in the number and the size of the air bubbles. Theperiphery of each air cell in fully whipped foams is composed primarily offat globules (5), but there is also a variable amount of original proteinadsorbed at the air–water interface. The rigidity of the foam structure isprovided by bridges of clustered fat particles connected to adjoining air cellsand by the formation of a well-developed fat particle structure in the bulkphase.

F. Film Tension, Elasticity, and Thickness Stability

The film elasticity and dynamic film stability of two types of fat-in-wateremulsions were investigated: a 12-wt% fat-in-water emulsion and a 20-wt%fat-in-water emulsion. Figures 11 and 12 show the data of foam film stressrelaxation with a relative film area expansion of 26%, 46%, and 66% for the20-wt% fat-in-water emulsion and the 12-wt% fat-in-water emulsion,respectively. Initially, the film was expanded very quickly; the surfactantconcentration on the film surface suddenly decreased due to the increasedfilm area. As a consequence, the film tension jumped from its equilibriumvalue to the maximum value. After that, because the concentration of sur-factants in the meniscus was higher than that of in the film, the surfactantsdiffused from the meniscus to the inside of the film, and from the inside ofthe film to the film surface; as a consequence, the film tension decreased andfinally reached the equilibrium film tension (in approximately 15min).Figures 11 and 12 show that the equilibrium values of the foam film tensionof 20wt% and 12wt% fat-in-water emulsions are both approximately91 dyn/cm. However, the initial value of the film tension right after theexpansion and the value of the initial slope of the curved film tensionversus time were different (Figs. 11 and 12). These data are presented inTable 1. Based on the data, the film elasticity was calculated for the 20-wt%fat and 12-wt% fat foam films formed from the corresponding emulsions(Fig. 13). The film elasticity data show that the 20-wt% fat foam film had a


Figure 12 Foam film stress relaxation with relative film expansion area: 26%,

46%, and 66% for fat-in-water emulsion (fat concentration: 20%).

Figure 11 Foam film stress relaxation with relative film expansion area: 26% and

46% for fat-in-water emulsion (fat concentration: 12wt%).


higher film elasticity, indicating that the 20-wt% fat foam film should have ahigher dynamic film stability. This showed a good correlation with thestability sequence of whipped foams produced from the correspondingfood emulsions (i.e., 20wt% fat whipped foam >12wt% whipped foam).

It is also found that the initial slope of the curves shown in Figs. 11and 12 for the 12-wt% fat-in-water emulsion was higher than the slope forthe 20-wt% fat-in-water emulsion, indicating that the diffusion rate of sur-factants (polysorbate 60) in a 20-wt% fat-in-water emulsion was slower thanthat of the 12-wt% fat-in-water emulsion. The Marangoni–Gibbs effecton the stability of foam films was more pronounced for the 20-wt%

Table 1 Foam Film Rheological Properties

Sample

Equilibrium

film tension

(mN/m)

Initial slope of the

curved film tension

versus time

(26% expansion)

(mN/m s)

Gibbs film

elasticity

(mN/m)

20wt% Fat-in-water

emulsion 91 �12.7 125.9

12wt% Fat-in-water

emulsion 92 �25.0 104.2

Figure 13 Foam film elasticity for fat-in-water emulsions.


fat-in-water emulsion sample, and the foam film produced from 20-wt%fat-in-water emulsion should be more stable. This shows a good correla-tion with the stability sequence of whipped foams produced from thecorresponding food emulsions.

Additionally, the film expansion ratios that the two systems could tole-rate without rupturing were different. No rupture for the 20-wt% fat systemwas observed in the course of the experiments for the expansion ratios of26% and 46%; the 12-wt% fat system was stable and ruptured only at 26%expansion. This film ruptured in about 3 s at a 46% expansion ratio. The20-wt% fat system withstood a 66% expansion ratio for 57 s. In this experi-ment, the initial foam film area was 1.4mm2 for all samples. The data showthe same conclusion as the elasticity measurement that the foam filmproduced from the 20-wt% fat-in-water emulsion—it had a higher dynamicstability than that produced from the 12-wt% fat-in-water emulsion.

We have observed a similar correlation between the foam film elasticityand the stability of the whipped dairy cream emulsion using polyglycerinfatty acid ester and sucrose ester as emulsifiers (35). It was found that atthe same mole of emulsifiers (3.3� 10�4M) the emulsion containing 30wt%fat and polyglycerin fatty acid ester had a film elasticity of 57.8 dyn/cmcompared to 46.5 dyn/cm for the emulsion containing sucrose ester, therebydemonstrating that greater stability results from a higher film elasticity.

G. Foam Film Air Permeability

The stability of aerated emulsions also depends on the air permeabilitythroughout the foam (32). Figure 14 presents the experimental foambubble diameter as a function of time for the 20-wt% fat-in-water emulsionand the 12-wt% fat-in-water emulsion. The initial foam bubble diameterwas 1mm for both samples. The rate of bubble shrinkage for foam lamellaformed from the 12-wt% fat-in-water emulsion was about 1.5 times fasterthan that of lamella formed from the 20-wt% fat-in-water emulsion. Similarobservations were made when the experimental capillary pressure, foam filmsurface area, and foam volume were plotted as a function of time.

The calculated foam lamella permeability/film thickness for the fat-in-water emulsion systems is presented in Fig. 15. The value of the foamlamella permeability/film thickness for the 12-wt% fat-in-water emulsionhas a maximum (15� 106 darcy/cm) at 150min. The data for the 20-wt%fat-in-water emulsion are almost in a straight line. At a film thickness of200 nm, the foam lamella formed from the 12-wt% fat-in-water emulsion at150min has an air permeability of about 50 darcy; under the same condi-tions, the foam lamella formed from the 20-wt% fat-in-water emulsion has apermeability of about 5 darcy (10 times lower). One can conclude that


Figure 14 Experimental foam bubble diameter versus time for fat-in-water

emulsions.

Figure 15 Foam lamella air permeability of fat-in-water emulsions.


smaller foam film permeability values result in greater stability in foam-based products.

H. Long-Term Stability of Model Aerated Food Emulsions

As discussed earlier, the dispersed air within the aerated food emulsions maydegrade because of Ostwald ripening and the gas-diffusion process. Theroles of these processes were studied using a model system (mesophase)before studying the effect of these processes on actual aerated food emulsionsystems. In analyzing the long-term stability of the aerated mesophase, it isimportant to measure the change in the overrun and the change in thebubble size distribution over time. It must also be noted that the contribu-tion of the large bubbles to the overrun is much more significant than thecontribution from the small bubbles because the volume of air in a bubble ofdiameter d varies as d3. Therefore, it is more instructive to interpret thestability of the aerated mesophase using a plot of the volume distributionof bubbles rather than the number distribution. The bubble volume at eachsize for the samples after a period of 1 week was multiplied by the ratio ofthe final to initial overrun and is plotted for the propylene glycol stearate–sucrose stearate system and triglycerol stearate–sucrose stearate system(Figs. 16 and 17, respectively). In order to present the loss of air as afunction of the bubble size distribution, the initial overrun was chosen asa benchmark. This kind of normalization accounts for the change in theoverrun and brings all curves to the same basis (viz. the volume of the totalsystem occupied by bubbles of different sizes). The volume distributionshifts to the large-size range and a significant fraction of the air is retainedin bubbles in the size range of 70 mm or more for the propylene glycolstearate system (Figs. 16 and 17). Note that bubbles in this size rangewere absent to begin with; even after a period of 1 week, the bubbles ofthis size are small in number, but significant in their contribution to the gasvolume. The practical implication of this effect is that this aerated meso-phase will be less suitable for use in the food emulsion because the largebubble size will tend to phase separate under the effect of buoyancy. Incomparison, the volume distribution of the triglycerol stearate systemdoes not change as much with time (i.e., after 1 week) and the bulk of theair volume is retained in bubbles of nearly the same size range—as was thecase initially. Second, it is to be noted that the change in the overrun over aperiod of 1 week is significantly more for the propylene glycol stearatesystem than it is for the triglycerol stearate system, implying that theformer is less stable than the latter. Finally, it is of significance that thestorage temperature does not affect the size distribution to a great extent,but affects the overrun. The change in the overrun is due to the diffusion of


gas from the bubbles to the atmosphere. The rate of diffusion is greaterat higher temperatures, resulting in the loss of more air from theaerated product and a lower overrun over a fixed period of time. Thereason for the difference in the degree of air loss to the atmosphere betweenthe two systems is due to the difference in the diffusivity of air in thetwo liquids.

1. Measurement of the Gas-Diffusion Rate

A bubble is formed inside the mesophase and the diameter of the bubble ismeasured as a function of time to determine the rate of air loss from a singlebubble; this would give an idea about the rate of gas loss throughthe bulk liquid medium. The pressure difference between the bubble andthe atmosphere is the main cause for gas transfer from the bubble to theatmosphere. The rate of gas loss depends on the size of the bubble, thesurface tension of the liquid, the permeability of the liquid film, the distance

Figure 16 Long-term stability results for the propylene glycol surfactant system.

The effect of temperature on volume distribution (corrected for overrun) of bubbles

is shown. The solid circle (f) represents results just after whipping (overrun 1000%);

the open circle (�) represents the results for the samples stored at 10�C for 1 week

(overrun 320%) and the solid square (g) represents the results for samples stored at

25�C for 1 week (overrun 220%).


from the free gas–liquid surface. Consider that a bubble of radius r0 isplaced inside the liquid at distance h from the gas–liquid interface.Pressure inside the bubble would be higher than the atmospheric pressuredue to the capillary effect. This pressure difference would cause the gas todiffuse from the bubble to the atmosphere. De Vries (22,23) has modeledthis phenomenon and related the radius of a shrinking bubble with timewhen the bubble is inside the liquid using the ideal gas law and Fick’s firstlaw [Eq. (12)].

The slopes were 4.4 and 6.7 mm2/min for the triglycerol stearate–sucrose stearate system and the propylene glycol stearate–sucrose stearatesystem, respectively, when the square of the radius was plotted as a functionof time and tile bubble was formed at a distance 250 mm from the freesurface. Results suggest that the diffusivity of the gas from the propyleneglycol stearate–sucrose stearate system is one and half times greater thanthat of the triglycerol stearate–sucrose stearate system. This explains thehigher overrun loss for the propylene glycol system.

Figure 17 Long-term stability results for the triglycerol stearate surfactant system.

The effect of temperature on volume distribution (corrected for overrun) of bubbles

is shown. The solid circle (f) represents results just after whipping (overrun 820%).

The open circle (�) represents the results for the samples stored at 10�C for 1 week

(overrun 450%) and the solid square (g) represents the results for samples stored at

25�C for 1 week (overrun 270%).


2. Ostwald Ripening Model and the Gas-Diffusion Model

A schematic of a section of aerated mesophase is shown in Fig. 18. Airbubbles are assumed to be surrounded by a layer of continuous mediumof thickness and at a distance hi from the free surface (i.e., the aeratedsystem is in contact with the atmosphere, which we can imagine as a bubbleof infinite radius). �mi, �si, and �1 are the volume fraction of the gas phasein a continuous medium, at the bubble surface, and at the free surface,respectively. The difference in the concentration of the gas phase is due tothe capillary pressure and the effect of the surrounding bubbles. The gasphase concentration in the continuous medium (�mi) can be assumed to bealways greater than that at the free surface (�1). This difference would leadto the transfer of gas from the continuous medium to the free surface. DeVries (22,23) has modeled gas transfer from a single bubble through a con-tinuos medium to a free surface. �mi would be greater than the �si forbubbles larger than the critical radius and less than the �si for bubblessmaller than the critical radius. [Yarranton and Masliyah (36) have modeledthe similar phenomena for the emulsion system]. Thus, air would be trans-ferred from the continuous medium to bubbles larger than the critical radiusand air would diffuse out of the bubbles smaller than the critical radius. Theexchange of gas among the bubbles is called Ostwald ripening.

a. Ostwald Ripening. The change in radius of a bubble due toOstwald ripening can be described by the following equations (36):

dr

dt¼

rðrþ ÞD

r2ð�mi � �siÞ ð13Þ

Figure 18 Schematic representation of the gas-diffusion and Ostwald ripening

processes.


where D is the gas diffusion coefficient, r is the radius of the bubble, t is thetime, and , �mi, �si are as defined earlier.

Yarranton and Masliyah (36) provided an expression for thedispersed-phase volume fraction in the continuous phase surrounding thedroplet as a function of the droplet diameter and droplet size distribution:

�mi ¼ �si þ��1

ðri þ Þ

1

r10�

1

ri

� �

ð14Þ

where r10 is the average bubble size (i.e., average radius of the bubbles) anda is defined as � ¼ 2vd=RT , where is the interfacial tension between thetwo phases, vd is the molar volume of the dispersed phase, R is a universalgas constant, and T is the temperature.

b. Gas Diffusion. Ostwald ripening does not predict any gas lossfrom the system. However, the overrun (i.e., amount of incorporatedair) of the system decreases with time because of gas diffusion to theatmosphere. If Fick’s first law is used to describe gas diffusion to the atmo-sphere from a bubble having radius ri and at a distance hi from the freesurface, the following equation is obtained:

dNA

dt¼

4�Dðþ riÞ2ð�mi � �1

hið15Þ

Now, because gas diffuses from the continuous medium surrounding thebubble, the dispersed phase volume will change:

��mi ¼�NAi

�Við16Þ

So, we can obtain

d�mi

dt¼

3Dðri þ Þ2ð�mi � �1Þ

hi½3riðþ riÞ þ 3�ð17Þ

Using Eq. (14), we obtain

d�mi

dt¼

d�sidt

þd

dt

��1

ðri þ Þ

1

r10�

1

ri

� ��

ð18Þ

d�mi

dt¼

dri

dt

��1r21

��1

ðri þ Þ2

2

riþ

r21�

1

r10

� ��

ð19Þ


Rearranging Eq. (19) and using Eq. (17), we can write

dri

dt¼ �gð,�1, ri, r10Þ

3Dðri þ Þ2ð�mi þ �1Þ


where the function gð,�1,ri,r10Þ is defined as

1

gð,�1,ri,r10Þ¼

��1r21

��1

ðri þ Þ2

2

riþ

r21�

1

r10

� ��

ð21Þ

Combining Eqs. (13) and (21), one can obtain the composite effect ofOstwald ripening and gas diffusion:

dri

dt¼

rðrþ ÞD

r2ð�mi � �siÞ � gð,�1,ri,r10Þ

3Dðri þ Þ2ð�mi � �1Þ


3. Model Prediction

We observed that overrun decreases substantially over 1 week for bothexperimental systems. The bubble size distribution also changes significantlyover the same period. Loss of gas could be attributed to the gas diffusion tothe atmosphere. In order to obtain quantitative agreement, we simulatedmodels for gas diffusion and the effect of Ostwald ripening. The simulationwas conducted on 120,000 bubbles (with an initial size distribution thatmatches the experimental one) in a cylindrical vessel 12 cm in height. Withthis number of bubbles, the system volume that corresponds to an airvolume overrun of 1000 vol% for propylene glycol system and 820 vol%for the triglycerol stearate system is a small fraction of the experimentalsystem volume, although the simulation cell has the same height as theexperimental system. This reduction in scale (i.e., smaller area) is necessaryin order to keep the number of bubbles in the simulation manageable and itwas found that this reduction in scale does not affect the results. For com-putational purposes, the cylinder was divided into 1200 layers. The thicknessof a layer (100 mm) was chosen to accommodate the largest air bubbleobserved in the experiment.

a. Effect of Gas Diffusion on Overrun. The effect of gas diffusion onoverrun (i.e., the long-term stability) was studied in two cases. In the firstcase, the De Vries model [Eq. (12)] was simulated for the bulk system. It wasassumed that the radius of each bubble decreases with time as predicted byEq. (12). The distance of each bubble from the free surface was known


because we had fixed the position of the each bubble in the vertical direc-tion. The rate [slope of Eq. (12)] was obtained from the shrinking bubbleexperiments. The results of our simulations for the system are presented inFig. 19. The De Vries (22,23) model predicts that the overrun of the trigly-cerol stearate–sucrose stearate system would decrease from 820% to 176%and the overrun of propylene glycol–sucrose stearate system would decreasefrom 1000% to 78%. However, for a finite dispersed-phase volume concen-tration, the gas-diffusion rate would be affected by the gas volume fractionand the bubble size distribution as described by Eq. (17). The simulationresults are shown in Fig. 19. We did not see significant improvement whencompared to the prediction of the De Vries (22,23) model. The results areunderstandable because our average bubble size was large (20 mm). Whenthe bubble size is large, the contribution of the second term of Eq. (13)becomes negligible. This would make Eq. (17) the same as Eq. (12); bothmodels predict almost the same decrease in the overrun for the system.When the average bubble size is small (as much as 5 mm), we observed asignificant difference (as much as 50%) between the predictions of the DeVries model and Eq. (20).

b. Effect of Ostwald Ripening on Overrun. We observed the theore-tical predictions were lower than the experimental observations when onlythe effect of gas diffusion was taken into account (overrun decreases from

Figure 19 Comparison of experimental and model predicted overruns after 1 week

storage at room temperature.


820% to 270% and 1000% to 220% for the triglycerol stearate and propy-lene glycol systems, respectively). The bubbles would transfer gas amongthemselves because of the capillary pressure difference. Large bubbles con-tinue to grow at the expense of smaller bubbles. The system approaches astate where the capillary pressure (being inversely proportional to the radius)in most of the bubbles is much lower than it was initially. Thus, the drivingforce for the transfer of air to the atmosphere decreases with time, resulting ina lower rate of air loss. It must be also be noted that in a system such as theone we studied experimentally, the distance between bubbles is very small,thereby also reducing the diffusion path for Ostwald ripening. Thus, one canexpect a reduction in the air loss from the system. We combined the effect ofgas diffusion and Ostwald ripening in Eq. (22). The value ofD�1� is taken as5� 10�4 mm3/s and 8� 10�4 mm3/s for the triglycerol stearate and propyleneglycol systems, respectively. The estimated value of D�1� for the nitrogen-in-water system is 2� 10�3mm3/s. The value for D�1� for the triglycerolsystem was chosen in order to match the final overrun. The value of D�1�for the propylene glycol system was estimated by using the ratio of the gas-diffusion rate for these two systems [8� 10�4

� 5� 10�4 (6.8/4.4)]. Figure 19shows that when the effects of gas diffusion and the Ostwald ripening processare taken into account, the model predictions are in good agreement with theexperimental results. Figure 20 depicts the effect of Ostwald ripening onthe long-term stability of the aerated mesophase. It can be concluded thatthe Ostwald ripening process retards the gas-diffusion process by transfer-ring the gas to larger bubbles (the gas-diffusion rate is inversely proportionalto the bubble radius).

The bubble size distribution (measured independently) was also usedto verify the model predictions. Figure 21 plots the experimental and modelpredictions of the volume of air distribution among bubbles of differentsizes on a cumulative basis for the triglycerol stearate system. The datacorresponding to 1 week of aging have been corrected to account for theloss of air from the system. The volume fraction of air in this case refers tothe ratio of the volume of air to the initial volume of the system, in order toaid comparison with the initial data. There is a reasonable match betweenthe model and experimental data for the bubble size distribution.

4. Application of Model in Predicting Long-Term Stability of an AeratedEmulsion System

The above model was used to predict the long-term stability of the twodifferent aerated emulsion systems (one containing fatty acid ester as theemulsifier and the other having the sucrose ester as the emulsifier). Thebubble size predicted using the model is compared with the experimentally


Figure 21 Comparison of experimental bubble size distribution with the model

prediction for triglycerol stearate system.

Figure 20 Effect of Ostwald ripening on gas diffusion.


observed bubble size after 1 week. After a period of 1 week, the mean bubblesize increased for both emulsion systems. The mean bubble size increasedfrom 41.2 mm to 60 mm in the case of the emulsion containing the fatty acidester. This confirms our earlier observation about the effect of Ostwaldripening on the bubble size distribution (i.e., larger bubbles grow at theexpense of smaller bubbles and this results in a higher mean bubble size).We also obtained good agreement between the experimentally observedbubble size and the model predicted bubble size for both systems (35).

IV. CONCLUSIONS

The study of fat particle structure during the whipping process such ashomogenization, cooling, aging, aeration, and whipping revealed that avery poor fat particle structure was produced immediately after homogeni-zation due to the very high temperature of the food emulsions. The fatparticle structure improved greatly after cooling. It was observed that thefat particle structure inside low-fat products (5wt% fat whipped foam and12wt% fat whipped foam) does not change with the aging time, whereas thefat particle structure is slightly improved for a high-fat product (20wt% fatwhipped foam), indicating that most of the adsorption of the surfactant andprotein on the fat particle surface have been completed before aging. The fatparticle structure improved greatly for all the products after whipping, indi-cating that a well-developed fat particle structure is formed in the bulk phaseand stabilizes the foam products.

The foam film rheology and foam film thickness stability producedfrom fat-in-water emulsions have been investigated using a film rheometer.Several important physical properties, such as equilibrium film tension,dynamic film tension, foam film elasticity, and critical film expansionarea, were obtained. A good correlation exists between the stability offoam-based products and the foam film rheological properties. A highfoam film elasticity and critical film expansion area and a small value ofthe initial slope of the curved film tension versus time led to more stablefoam-based products.

The stability of aerated emulsions also depends on the air permeabilitythrough the foam lamella. Smaller values for foam film permeability yield ahigher stability in foam-based food products.

From numerical simulations and experimental studies of the destabi-lization of the aerated mesophase (overruns were 820% and 1000% for thetriglycerol system and the propylene glycol system, respectively), we foundthat the predominant mechanisms of destabilization are gas diffusion to theatmosphere and Ostwald ripening. The overrun of the system decreases


(from 820% to 270% and from 1000% to 220% for the triglycerol systemand propylene glycol system, respectively) with time because of the gasdiffusion from the bubbles to the atmosphere. The rate of gas diffusiondepends on the permeability of air through the adsorbed layer and meso-phase and on the bubble size (driving force is inversely proportional to thebubble size).

Bubbles would transfer gas among themselves due to the capillarypressure difference (the Ostwald ripening phenomenon). Large bubbles con-tinue to grow at the expense of smaller bubbles. Thus, the driving force(being inversely proportional to the radius) for the transfer of air to theatmosphere decreases with time, resulting in a lower rate of air loss. Ostwaldripening also affects the bubble size distribution. Bubbles smaller than thecritical size shrink and may eventually disappear. A theoretical model wasdeveloped to predict the effects of gas diffusion and Ostwald ripening on thelong-term stability of the aerated food products. The gas-diffusion rate wasobtained from the shrinking bubble experiments. The model was verifiedwith the experimental results of overrun and good agreement was observed.Model predictions for the bubble size distribution as a function of time arefound to be in agreement with the experimentally observed bubble sizes,which were determined by conducting an independent set of experiments.

REFERENCES

1. W. Buchheim, N. M. Barlord, and N. Krog, Relation between microstructure

destabilization phenomena and rheological properties of whippable emulsions,

Food Microstruct. 4, 221–232 (1985).

2. E. C. Needs, Importance of mile proteins to the whippable properties

of 38% fat cream, in Food Colloids, The Proceedings of an International

Symposium Organized by the Food Chemistry Group of the Royal Society of

Chemistry, 1988.

3. E. C. Needs and B. E. Brooker, Factors affecting the formation and stabilization

of high fat foams, in Food Polymers, Gels and Colloids, The Proceedings of an

International Symposium Organized by the Food Chemistry Group of the Royal

Society of Chemistry, 1990, pp. 227–231.

4. D. F. Darling, Recent advances in the destabilization of dairy emulsions, J. Dairy

Res. 49, 695–712 (1982).

5. B. E. Brooker, M. Anderson, and A. T. Andrews, The development of structure

in whipped cream, Food Microstruct. 5, 277–285 (1986).

6. B. E. Brooker, The adsorption of crystalline fat to the air–water interface of

whipped cream, Food Struct. 9, 223–229 (1990).

7. W. Xu, A. Nikolov, D. T. Wasan, A. Gonsalves, and R. P. Borwankar, Particle

structure and stability of colloidal dispersions as probed by Kossel diffraction

technique, J. Colloid Interf. Sci. 191, 471–481 (1997).


8. E. Dickenson, Food colloids—An overview, Colloids Surfaces 42, 191–204

(1989).

9. J. Courthandon, E. Dickinson, and D. G. Dalgleish, Competitive adsorption of

b-casein and nonionic surfactants in oil-in-water emulsions, J. Colloid Interfa.

Sci. 145(2), 391–395 (1991).

10. L. A. Lobo, Stability of emulsions and foams, PhD thesis, Illinois Institute of

Technology, Chicago, 1991.

11. E. Dickinson, Interfacial interactions and the stability of oil-in-water

emulsions, Pure Appl. Chem. 64(11), 1721–1724 (1992).

12. P. Richmond, Food technology and nutrition: Challenges for colloid and

interface science, Pure Appl. Chem. 64(11), 1751–1755 (1992).

13. J. Chen, E. Dickinson, and G. Iveson, Interfacial interactions, competitive

adsorption, and emulsion stability, Food Struct. 12, 135–146 (1993).

14. J. M. Soos, K. Koczo, E. Erdos, and D. T. Wasan, An automatic apparatus for

measuring interfacial and film tension under static and dynamic conditions,

Rev. Sci. Instrum. 65(11), 3555–3562 (1994).

15. Y. H. Kim, D. T. Wasan, and P. J. Breen, A study of dynamic interfacial

mechanisms for demulsification of water-in-oil emulsions, Colloids Surfaces 95,

235–247 (1995).

16. Y. H. Kim, A. D. Nikolov, D. T. Wasan, H. Diaz-Arauzo, and C. S. Shetty,

Demulsification of water-in-crude oil emulsions: Effects of film tension,

elasticity, diffusive, and interfacial acitivity of demulsified individual compo-

nents and their blends, J. Dispers. Sci. Tech. 17(1), 33–53 (1996).

17. Y. H. Kim and D. T. Wasan, Effect of demulsifier partitioning on the destabi-

lization of water-in-oil emulsions, Ind. Eng. Chem. Res. 35(4), 1141–1149 (1996).

18. Y. H. Kim, K. Koczo, and D. T. Wasan, Dynamic film and interfacial tensions

in emulsion and foam systems, J. Colloid Interfa. Sci. 187, 29–44 (1997).

19. B. Sood, Study of foam dynamics, texture and stability with application to

aerated and foamed products, MS thesis, Illinois Institute of Technology,

Chicago, pp. 2.1–2.12.

20. I. M. Lifshitz and V. V. Slyozov, The kinetics of precipitation from

supersaturated solid solutions, J. Phys. Chem. Solids 19(1/2), 35 (1961).

21. C. Wagner, Ber. Bunsenges. Phys. Chem. 65, 581 (1961).

22. A. J. De Vries, Foam Stability Part I, J. Recueil 77, 81–91 (1958).

23. A. J. De Vries, Foam Stability Part II, J. Recueil 77, 209–223 (1958).

24. H. M. Princen and S. G. Mason, The permeability of soap films to gases.

J. Colloid Sci. 20, 353–375 (1965).

25. H. M. Princen, J. T. G. Overbeek, and S. G. Mason, The permeability of soap

films to gases. II. A simple mechanism to monolayer permeability. J. Colloid

Sci. 24, 125–130 (1967).

26. A. Prins, Theory and practice of formation and stability of food foams, in Food

Emulsions and Foams (E. Dickenson, ed.), Special Publicaiton No. 58, The

Royal Society of Chemistry, London, 1987, pp. 30–39.

27. A. Arefmanesh and S. G. Advani, Rheol. Acta 30, 274 (1991).

28. D. C. Venerus and N. Yala, Aiche J. 43, 2948 (1997).


29. D. C. Venerus, N. Yala, and B. Bernstein, J. Non-Newtonian Fluid Mech. 75,

55 (1998).

30. D. T. Wasan and A. D. Nikolov, Structure and stability of emulsions, in

Encyclopedic Handbook of Emulsion Technology (J. Sjoblom, ed.), Marcel

Dekker, New York, 2001, pp. 59–70.

31. R. Nagargan, K. Koczo, E. Erods, and D. T. Wasan, Controlled drop

tensiometer for measuring dynamic interfacial tension and film tension, AICE

J. 41(4), 915–923 (1995).

32. M. E. Wijnen and A. Prins, Disproportion in aerosol whipped cream, in

Food Macromolecules and Colloids, (E. Dickenson and D. Lorient, eds.),

Special Publication No. 156, The Royal Society of Chemistry, London, 1995,

pp. 309–311.

33. F. Dullien, Porous Media Fluid Transport and Pore Structure, Academic Press,

New York, 1979.

34. H. D. Goff, The role of chemical emulsifiers and dairy proteins in fat

destabilization during the manufacture of ice cream, PhD thesis, Cornell

University, Ithaca, NY, 1988.

35. S. Uchil, Structure and stability of imitation dairy cream emulsion, MS thesis,

Illinois Institute of Technology, Chicago, 2001.

36. H. W. Yarranton and J. H. Masliyah, Numerical simulation of ripening in

emulsions, J. Colloid Interfa. Sci. 196, 157–169 (1997).


7Surface Forces and EmulsionStability

Per M. Claesson, Eva Blomberg, and Evgeni PoptoshevRoyal Institute of Technology, Stockholm, Sweden

I. INTRODUCTION

Emulsions are dispersions of one liquid in another liquid, most commonlywater-in-oil or oil-in-water. The total interfacial area in an emulsion is verylarge, and because the interfacial area is associated with a positive freeenergy (the interfacial tension), the emulsion system is thermodynamicallyunstable. Nevertheless, it is possible to make emulsions with an excellentlong-term stability. This requires the use of emulsifiers that accumulate atthe oil–water interface and create an energy barrier toward flocculation andcoalescence. The emulsifiers can be ionic, zwitterionic, or nonionic surfac-tants, proteins, amphiphilic polymers, or combinations of polymers andsurfactants. The structure of the adsorbed layer at the water–oil interfacemay be rather complex, involving several species adsorbed directly to theinterface as well as other species adsorbing on top of the first layer.

The first question one may ask is if an oil-in-water emulsion or awater-in-oil emulsion is formed then the two solvents are dispersed intoeach other with the use of a given emulsifier. There are several empiricalroles addressing this problem. The first is due to Bancroft (1), who statedthat if the emulsifier is most soluble in the water phase, then an oil-in-wateremulsion will be formed. A water-in-oil emulsion will be obtained when thereverse is true. The HLB (hydrophilic–lipophilic balance) concept is used fordescribing the nature of the surfactant. It was first introduced by Griffin (2)and later extended by Davies (3). Hydrophobic emulsifiers having alow HLB number, say below 6, are predicted to be suitable for forming


water-in-oil emulsions, whereas more hydrophilic emulsifiers with high HLBvalues, above about 10, are suggested to be suitable for forming oil-in-wateremulsions. The HLB value can easily be calculated from the structure of theemulsifier (3). An HLB value has also been assigned for most common oils.It is defined as the HLB number of the emulsifier in a homologous seriesthat produces the most stable oil-in-water emulsion. A nonpolar oil is foundto have a lower HLB number than a polar oil. Hence, the choice ofemulsifier has to be adjusted to the type of oil that is to be emulsified.

The use of an HLB value for nonionic emulsifiers of the oligo (ethyleneoxide) type has its drawbacks because their properties are strongly tempera-ture dependent. This is clearly seen in three-component oil–water–surfactantphase diagrams. At low temperatures, microemulsions of oil droplets inwater (Winsor I) are formed. In a small temperature interval, bicontinuousmicroemulsions (Winsor III) are stable, followed at higher temperatures by amicroemulsion consisting of water droplets in oil (Winsor II). These transi-tions are due to a change in the spontaneous monolayer curvature frompositive at low temperatures to negative at high temperatures. This behavioris closely mimicked by the thermodynamically unstable (macro)emulsions,and it is common to describe these emulsions in terms of the phase-inversiontemperature (PIT). Below the PIT, the emulsion is of the oil-in-water type,whereas above the PIT, it is of the water-in-oil type. Very close to the PIT, nostable (macro)emulsions can be formed. It has been argued that this changein behavior, as for the microemulsions, is due to the change in spontaneouscurvature of such surfactant films at the oil–water interface, particularly theease with which hole formation leading to coalescence occurs (4). Note thatthe PIT depends not only on the nature of the emulsifier but also on the typeof oil used, which often can be explained by the degree of oil penetration intothe emulsifier film.

When two droplets approach each other, they will interact with hydro-dynamic forces and with surface forces of molecular origin. Finally, when thedroplets are close enough, they may coalesce and form one larger droplet. Anemulsion will have a long-term stability if the droplets are prevented fromcoming close to each other by strong repulsive forces and if they are pre-vented from coalescing even when they are close to each other. However, inthis case also, a slow destabilization due to Ostwald ripening will occur.

II. INTERACTIONS AND HOLE FORMATION

In this section, we will give a short overview of hydrodynamic and surfaceforces as well as hole formation leading to coalescence. References will beprovided for the reader who wants to delve further into these subjects.


A. Hydrodynamic Interactions

When liquid drains from the gap between two approaching spherical emul-sion droplets of equal size, a hydrodynamic force resulting from viscousdissipation is produced. As long as the surfaces do not deform (i.e., smallforces) and the liquid next to the surface is stationary (no slip condition, seenext page) the hydrodynamic force is given by (5).

F ¼ �3��R2

2D

dD

dtð1Þ

where R is the radius of the spheres, � is the viscosity of the draining liquid,D is the separation between the spheres, and t is the time. This equationdescribes the hydrodynamic interaction when the droplets are far apart anddo not interact with each other very strongly. However, as soon as theinteraction between the surfaces is sufficiently large, the emulsion dropletswill deform and Eq. (1) is no longer valid.

In concentrated emulsions, we meet another extreme case. A thinplanar liquid film now separates the emulsion droplets, and they maychange their shape from spherical to polyhedral (6). In this case, theliquid drains out of the flat part of the film owing to the capillary suctionpressure. The outflow of liquid between rigid parallel disks was consideredby Reynolds (7,8) who found that the pressure varied with the radialdistance from the center of the disk as

P ¼ P0 þ3�

D3ðr20 � r2ÞVR ð2Þ

where P is the pressure at a distance r from the center, r0 is the radius ofthe plate, P0 is the hydrostatic pressure, which equals the total pressure atthe edge of the contact, (i.e., at r¼ r0, and VR is the rate of approach(i.e.,�dD/dt).

The repulsive hydrodynamic force acting on the plates is obtained byintegrating over the plate area and subtracting the hydrostatic pressurecontribution:

F ¼ 2�

Z r0

0

ðP� P0Þr dr ¼ 2�

Z r0

0

3�

D3ðr20 � r2ÞVr

� �

r dr

¼3�r40�VR

2D3ð3Þ


The average excess pressure (which equals the capillary pressure) betweencircular plates can be expressed as

�PP ¼F

�r20ð4Þ

Hence, we obtain the well-known Reynolds equation

VR ¼ �dD

dt¼

2D3 �PP

3�r20ð5Þ

We immediately see that the film-thinning rate is reduced, and thus theemulsion stability increased, by an increase in bulk viscosity. In the casein which the liquid film is so thin that surface forces no longer can beneglected, the capillary pressure term in the Reynolds equation should bereplaced by the total driving force (�P) for the thinning. This is equal to thedifference between the capillary pressure and the disjoining pressure (�) dueto the surface forces acting between the emulsion droplet surfaces: �P ¼

ð �PP��Þ: Clearly, a positive disjoining pressure (i.e., a repulsive force)reduces the driving force for film thinning and thus the drainage rate.

Experimentally determined rates of thinning do not always agree withthe predictions of the Reynolds model. For foam films stabilized by ananionic surfactant, sodium dodecyl sulfate (9,10), it has been shown thattypical thinning rates exhibited a much weaker dependence on the filmradius (r�0.8�0.9) than the predicted r�2 dependence. To obtain an under-standing for why the Reynolds theory of thinning does not always agreewith experimental results, it is worthwhile to consider two assumptionsmade when arriving at Eq. (5). First the result is valid only under ‘‘no-slip’’ conditions (i.e., the velocity of the liquid at the film interface isassumed to be zero). This is the case when the drainage takes place betweensolid hydrophilic surfaces. In contrast, only the adsorbed emulsifier layerprovides the surface rigidity in foam and emulsion films, and it is notobvious that the no-slip condition is fulfilled. The drainage rate would belarger than predicted by Eq. (5) if this condition was not valid. Jeelani andHartland (11), who calculated the liquid velocity at the interfaces of emul-sion films for numerous systems studied experimentally, addressed thispoint. They showed that even at a low surfactant concentration, the liquidmobility at the interface is dramatically reduced by the adsorbed surfactant.Hence, it is plausible that when the adsoption density of the emulsifier islarge (nearly saturated monolayers), the surface viscosity is high enough tovalidate the no-slip condition. It has been pointed out that a nonzero liquidviscosity at the interface is not expected to have an influence on the func-tional dependence of the drainage rates on the film radius (9). Hence, thedeviations found experimentally have to have another origin.


A second assumption made when arriving at Eq. (5) is that thedrainage takes place between parallel surfaces. Experimental studies onliquid films (9,10) have shown that during the thinning process, it iscommon that nonuniform films are formed which have a thicker region, adimple, in the center. For larger films, even more complicated, multidimpledprofiles have been found. Calculating the drainage rate for interfaces withsuch a complex shape is far from easy. However, recently Manev et al. (9)proposed a model for the drainage between nonparallel, immobile surfaces.The following expression has been proposed for the rate of thinning:

V ¼8

3�

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi4D12�P8

�121 3r4

5

s

ð6Þ

Here, �1 is the first root of the first-order Bessel function of the first kindand is the surface tension. Note that in Eq. (6), the rate of thinningis inversely proportional to r4/5. This is in good agreement with someexperimental observations.

III. SURFACE FORCES

At sufficiently small droplet separations, say below 100 nm, surface forceshave to be considered. These forces affect the drainage rate as well as theequilibrium interactions, particularly if flocculation occurs. The most com-monly encountered forces are briefly described in this section. For a generalreference to surface force, see the book by Israelachvili (12).

A. Van der Waals Forces

Van der Waals forces originate from the motion of negatively charged elec-trons around the positively charged atomic nucleus. For condensed materi-als (liquids or solids), this electron motion gives rise to a fluctuatingelectromagnetic field that extends beyond the surface of the material.Thus, when, for example, two particles or emulsion droplets are closetogether, the fluctuating fields associated with them will interact with eachother. The energy of interaction per unit area (Wvdw) between two equalspheres with radius R a distance D apart is given by:

Wvdw ¼�A

6

2R2

D2 þ 4RDþ

2R2

D2 þ 4RDþR2

�

þ lnD2 þ 4RD

D2 þ 4RDþR2

� ��

ð7Þ


where A is the nonretarded Hamaker constant. When the particle radius ismuch larger than the separation of the particles, Eq. (7) is reduced to

Wvdw ¼ �AR

12Dð8Þ

The Hamaker constant depends on the dielectric properties of the two inter-acting particles and the intervening medium. When these properties areknown, one can calculate the Hamaker constant. An approximate equationfor two identical particles (subscript 1) interacting across a medium (sub-script 2) is:

A ¼3kT

4

"1 � "2"1 � "2

� �2

þ3h�

16ffiffiffi2

pðn21 � n22Þ

2

ðn21 þ n22Þ3=2

!

ð9Þ

where k is the Boltzmann constant, T is the absolute temperature, � is themain absorption frequency in the ultraviolet (UV) region (often about3� 1015Hz), h is Planck’s constant, " is the static dielectric constant, andn is the refractive index in visible light.

From Eqs. (8) and (9), it is clear that the van der Waals interactionbetween two identical particles or emulsion droplets is always attractive.One may also note that the Hamaker constant for two oil droplets inter-acting across water is identical to the Hamaker constant for two waterdroplets interacting across oil.

B. Electrostatic Double-Layer Forces

Electrostatic double-layer forces are always present between charged parti-cles or emulsion droplets in electrolyte solutions. Counterions to the emul-sion droplet (ions with opposite charges to that of the drop) are attracted tothe surfaces and co-ions are repelled. Hence, outside the charged emulsiondroplet, in the so-called diffuse layer, the concentration of ions will bedifferent than that in bulk solution, and the charge in the diffuse layerbalances the surface charge.

An electrostatic double-layer interaction arises when two chargeddroplets are so close together that their diffuse layers overlap. The electro-static double-layer interaction, Wdl, for two identical charged drops with asmall electrostatic surface potential and a radius large compared to theirseparation is approximately given by

Wdl ¼ 2�R""0�20 expð��DÞ ð10Þ


where "0 is the permittivity of vacuum, " is the static dielectric constant ofthe medium, �0 is the surface potential, and ��1 is the Debye screeninglength given by:

��1 ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi"0"kT

1000NAe2P

i ciz2i

s

ð11Þ

where e is the charge of the proton, NA is Avogadro’s number, ci is theconcentration of ion i expressed as mol/dm3, and zi is the valeny of ion i.

The double-layer interaction is repulsive and it decays exponentiallywith surface separation with a decay length equal to the Debye length.Further, the Debye length and, consequently, the range of the double-layer force decreases with increasing salt concentration and the valency ofthe ions present. The famous Derjaguin Landau Verwey Overbeek (DLVO)theory for colloidal stability (13,14) takes into account double-layer forcesand van der Waals forces.

C. Hydration and Steric-Protrusion Forces

Hydration and steric-protrusion forces are repulsive forces that have beenfound to be present at rather short separations between hydrophilic surfacessuch as surfactant head groups. At least two molecular reasons for theseforces have been identified. First, when two polar surfaces are brought closetogether, the polar groups will be partly dehydrated, which gives rise to arepulsive force (15). Second, as two surfaces are brought close together, themolecules at the interface will have a decreased mobility perpendicular tothe surface, which decreases the entropy of the system, and this gives rise toa steric type of repulsion (16). Empirically, it has been found that thehydration/steric repulsion between surfactant and lipid head groupsdecays roughly exponentially with distance:

Whyd ¼ W0hyd exp �

D

� �

ð12Þ

where is the decay length of the force, typically 0.2–0.3 nm.

D. Polymer-Induced Forces

The presence of polymers on surfaces gives rise to additional forces thatcan be repulsive or attractive. Under conditions when the polymer isfirmly anchored to the surface and the surface coverage is large, a steric


repulsion is expected. As the surfaces are brought together, the segmentdensity between them increases, which results in an increased number ofsegment–segment contacts and a loss of conformational entropy of thepolymer chains. The conformational entropy loss always results in a repul-sive force contribution that dominates at small separations. The increasednumber of segment–segment contacts may give rise to an attractive or arepulsive force contribution. This is often discussed in terms of thechi-parameter (�-parameter) or in terms of the solvent quality. Undersufficiently poor solvent conditions (�>1/2), when the segment–segmentinteraction is sufficiently favourable compared to the segment–solventinteraction, the long-range interaction can be attractive. Otherwise, itis repulsive. The steric force can be calculated by using lattice meanfield theory (17) or scaling theory (18). The actual force encountered ishighly dependent on the adsorption density, the surface affinity, thepolymer architecture, and the solvency condition. Hence, no simple equa-tion can describe all situations. However, a high-density polymer layer, a‘‘brush’’ layer, in a good solvent, provides good steric stabilization. Thescaling approach provides us with a simple formula that often describesthe measured interactions under such conditions rather well (19). Itstates that the pressure P(D) between two flat polymer-coated surfaces isgiven by

PðDÞ �kT

s3D�

D

� �9=4

�D

D�

� �3=4" #

ð13Þ

where Eq. (13) is valid provided that the separation, D, is less than D�

(where D� is twice the length of the polymer tail), and s is the linear distancebetween the anchored chains on the surface. For the interactions betweentwo spheres with a radius significantly larger than their separation, thisrelation is modified to

FðDÞ

R� ��

Z D

D�

PðD0Þ dD0 ð14Þ

The parameters needed to calculate the force are the length of the extendedpolymer chain and the separation between the polymer chains on thesurface. The latter parameter can be estimated from the adsorbedamount, whereas the length of the polymer chains enters as a fittingparameter. The formula predicts a repulsion that increases monotonicallywith decreasing separation.


E. Coalescence and Hole Formation

When studying drainage and equilibrium interactions in single-foam filmsabove the critical micellar concentration (cmc) of the surfactant, it is oftenfound that the film thickness undergoes sudden changes (20,21). This phe-nomenon is known as stratification. Below the cmc, one sudden changefrom a water-rich common black film to a very thin Newton black filmmay occur. This transition does not occur uniformly over the whole filmarea but initially in some small regions. The thinner regions are often calledblack spots because they appear darker than the rest of the film when viewedin reflected light. Once formed, the size of a black spot grows as the liquiddrains out from the foam lamellae. Bergeron and co-workers noted that theviscous resistance to the flow in the thin film is large and that this leads to anincrease in the local film thickness next to the black spot (22,23). The sug-gested shape of the thin liquid layer, which is supported by experimentalobservations and theoretical calculations (22,23) is illustrated in Fig. 1. Inmany cases, no or unstable Newton black films are formed. In these cases,the films rupture due to formation of a hole that rapidly grows as a result ofsurface-tension forces. Emulsion coalescence occurs in a similar manner.

The mechanism of black spot formation and rupture has been exten-sively studied (24). It is generally recognized that the liquid film is unstablein regions of the disjointing pressure (�) isotherm (force curve) where thederivative with respect to film thickness (D) is larger than zero (i.e.,d�/dD>0). Hence, close to a maximum in the disjoining pressure isotherm(see Fig. 2), a small disturbance causing a change in film thickness and/orcapillary pressure may spontaneously grow and lead to significant change infilm thickness (e.g., Newton black film formation or rupture).

The stability of foams and emulsions depends critically on whetherformation of a stable Newton black film or a hole leading to coalescenceis favored. Kabalnov and Wennerstrom (4) addressed this question by

Figure 1 Illustration of shape of the thin liquid film around the position of a newly

formed black spot.


developing a temperature-induced hole nucleation model. They point outthat the coalescene energy barrier is strongly affected by the spontaneousmonolayer curvature. The authors consider a flat emulsion film, covered bya saturated surfactant monolayer, in thermodynamic equilibrium with amicellar bulk solution. The emulsion breaks if an induced hole growsalong the film having a thickness h¼ 2b (Fig. 3). The change in freeenergy occurring when a hole is formed is given as the difference in theinterfacial tension integrals over the interface for a film with a hole com-pared to that for a planar film:

W ¼

Zðx, y, zÞ dA

ðfilm with holeÞ

�

Zðx, y, zÞ dA

ðflat filmÞ

ð15Þ

The diving force for the formation of a hole is the reduction in freeenergy owing to a decrease in surface area of the planar part of the film,whereas it is counteracted by the increased free energy due to the surfacearea created around the hole. In general, the change in free energy goesthrough a maximum as the hole radius increases. One new feature of theKabalnov–Wennerstrom model is that the surface tension at the hole edge is

Figure 2 Typical disjoining pressure isotherm showing one maximum (A) and one

minimum (B); the film is unstable between points A and B.


considered to be different than that at the planar film surface. The reasonfor this is that the curvature of the interface is different, leading to a differ-ence in surfactant monolayer bending energy. This can be expressed as (4)

curved ¼ planar þ 2�ðH �H0Þ2� 2�H2

0 ð16Þ

Here, H and H0 are the mean and the spontaneous curvatures and � is thebending modulus.

Clearly, the surface tension has a minimum when the spontaneouscurvature of the surfactant film equals the mean curvature of the interface.The mean curvature for a flat interface is zero, larger than zero for aninterface curving toward the oil (oil-in-water emulsions), and smaller thanzero for a water-in-oil emulsion. Hence, a large positive spontaneous mono-layer curvature, as for a strongly hydrophilic surfactant, favors oil-in-wateremulsions and vice versa. The Kabalnov–Wennerstrom model also allowsthe thickness of the film to vary in order to minimize the free energy of holeformation (i.e., the mean curvature of the film close to the hole canapproach the spontaneous monolayer curvature). The Kabalnov–Wennerstrom model has to be solved numerically in order to calculate thecoalescence activation energy. However, a ‘‘big hole approach’’ where a� b(see Fig. 3) gives surprisingly good results. In this model, the energy forcreating a hole with radius a is given as

W ¼ 2�a� � 2�a2 ð17Þ

where 2�a is the circumference of the hole, g is the line tension, �a2 is thearea of the hole at each interface, and is the surface tension. The secondterm is the free energy gain by reducing the flat area of the film, and the firstterm is the energy penalty of creating the inside of the hole. The value of the

Figure 3 Geometry of the thin film just after a hole has been created. (Redrawn

from Ref. 4, with permission.)


line tension can be calculated when the spontaneous monolayer curvatureand the monolayer bending modulus is known (4). The activation energy ofcoalescence (W�) is obtained by finding the point where dw/da¼ 0, whichgives the final expression:

W� ¼��2

2ð18Þ

The particular feature with ethylene oxide-based surfactants is thattheir interaction with water is less favourable at higher temperatures. Thisleads to a decrease in the spontaneous monolayer curvature with tempera-ture, explaining the transition from oil-in-water emulsions below the PIT towater-in-oil emulsion above the PIT. In the vicinity of the PIT, the energybarrier against coalescence (W�) varies very strongly with temperature. Forthe system n-octane–C12E5–water, the following approximate relation wasobtained in terms of �T¼T�Td, where Td is the PIT (4):

W�

kT¼ 0:43þ 30:9j�T j ð19Þ

The predicted very steep increase in the coalescence barrier away from thePIT is qualitatively in good agreement with the experimentally observedmacroemulsion behavior (25).

IV. SURFACE FORCE TECHNIQUES

There are several methods available for measuring forces between two solidsurfaces, two particles, or liquid interfaces (26). In this section, we brieflymention some of the features of the techniques that have been used in orderto produce the results described in the later part of this chapter. The forcesacting between two solid surfaces were measured either with the inter-ferometric surface force apparatus (SFA) or with the MASIF (measure-ments and analysis of surface and interfacial forces). Interactions betweenfluid interfaces were determined using various versions of the thin-filmpressure balance (TFB).

A. Interferometric Surface Force Apparatus

The forces acting between two molecularly smooth surfaces, normally micaor modified mica, can be measured as a function of their absolute separationwith the interferometric SFA (Fig. 4) (27). This provides a convenient wayto measure not only long-range forces but also the thickness of adsorbed


layers. The absolute separation is determined interferometrically to within0.1–0.2 nm by using fringes of equal chromatic order. The surfaces are gluedon optically polished silica disks and mounted in the SFA in a crossedcylinder configuration. The surface separation is controlled either by adjust-ing the voltage applied to a piezoelectric crystal rigidly attached to one ofthe surfaces or by a synchronous motor linked by a cantilever spring to theother surface. The deflection of the force-measuring spring is also deter-mined interferometrically, and the force is calculated from Hooke’s law.For further details, see Ref. 27.

When an attractive force component is present, the gradient of theforce with respect to the surface separation, @F=@D, may at some distancebecome larger than the spring constant, k. The mechanical system thenenters an unstable region causing the surfaces to jump to the next stablepoint (compare instabilities in free liquid films that occur when d�=dD > 0).The adhesion force, F(0), normalized by the local mean radius of curvature,

Figure 4 Schematic of a surface force apparatus (SFA). The measuring chamber is

made from stainless steel. One of the surfaces is mounted on a piezoelectric tube that

is used to change the surface separation; the other surface is mounted on the force

measuring spring. (From Ref. 26, with permission.)


R, is determined by separating the surfaces from adhesive contact. The forceis calculated from

Fð0Þ

R¼

kDj þ FðDjÞ

Rð20Þ

where F(Dj) is the force at the distance (Dj) to which the surfaces jumped onseparation, and R is the mean radius of the surfaces.

B. The Bimorph Surface Force Apparatus

The force as a function of the surface separation between glass substratesurfaces was measured with a MASIF instrument (28). This apparatus isbased on a bimorph force sensor to which one of the surfaces is mounted(Fig. 5). The other surface is mounted on a piezoelectric tube. The bimorph(enclosed in a Teflon sheath) is mounted inside a small measuring chamber,which is clamped to a translation stage that serves to control the coarseposition of the piezoelectric tube and the upper surface.

The voltage across the piezoelectric tube is varied continuously and thesurfaces are first pushed together and then separated. The bimorph willdeflect when a force is experienced and this generates a charge in proportionto the deflection. From the deflection and the spring constant, the forcefollows simply from Hooke’s law. The motion of the piezo is measuredduring each force run with a linear displacement sensor. The signal togetherwith the signal from the bimorph charge amplifier, the voltage applied to thepiezoelectric tube, and the time are recorded by a computer. The speed ofapproach, the number of data points, and other experimental variables caneasily be controlled with the computer software.

When the surfaces are in contact, the motion of the piezoelectric tubeis transmitted directly to the force sensor. This results in a linear increase ofthe force sensor signal with the expansion of the piezoelectric tube. Thesensitivity of the force sensor can be calibrated from this straight line, andthis measuring procedure allows the determination of forces as a function ofseparation from a hard wall contact with a high precision (within 1–2 A indistance resolution). Note, however, that the assumption of a ‘‘hard wall’’contact is not always correct (29).

The MASIF instrument does not use interferometry to determinesurface separations, which leads to the drawback that the layer thicknesscannot be determined, but to the advantage that the instrument can be usedwith any type of hard, smooth surfaces. In most cases, spherical glasssurfaces are used. They are prepared by melting a 2-mm-diameter glassrod until a molten droplet with a radius of 2mm is formed.


C. Derjaguin Approximation

The force measured between crossed cylinders (Fc), as in the SFA, andbetween spheres (Fs), as in the MASIF, a distance D apart is normalizedby the local geometric mean radius (R). This quantity is related to the freeenergy of interaction per unit area between flat surfaces (W) according tothe Derjaguin approximation (30):

Fc

2�R¼

Fs

�R¼ W ð21Þ

This approximation is valid when the radius (about 2 cm in the SFA and2mm in the MASIF) is much larger than the surface separation (typically10�5 cm or less), a requirement fulfilled in these experiments. With the SFA,

Figure 5 Schematic of the MASIF instrument. The upper surface is mounted on a

piezoceramic actuator that is used for changing the surface separation; the hysteresis

of the piezoexpansion/contraction cycle can be accounted for by using a linear

variable displacement transducer (LVDT). The lower surface is mounted on a

bimorph force sensor. (From Ref. 26, with permission.)


the local radius is determined from the shape of the standing-wave pattern,whereas in the MASIF, we have used the assumption that the local radius isequal to the macroscopic radius, determined using a micrometer. The radiusused in Eq. (21) is that of the undeformed surfaces. However, under theaction of strongly repulsive or attractive forces, the surfaces will deform andflatten (31,32). This changes the local radius and invalidates Eq. (21)because R becomes a function of D.

D. Thin-Film Pressure Balance

Accurate information about the rate of thinning, the critical thickness ofrupture, and the forces acting between two air–water interfaces, betweentwo oil–water interfaces, and between one air–water and one oil–water inter-face can be gained by using thin-film pressure balance techniques. Thethickness of the separating water film is determined by measuring the inten-sity of reflected white light from a small flat portion of the film (33). Due tointerference of the light reflected from the upper and lower film surfaces,characteristic interference colors are observed during the thinning. Thesecolors correspond to a shift in the wavelengths undergoing constructiveand destructive interference. When such a process is recorded (normallyas intensity of a given light wavelength versus time), a sequence of intensityminima and maxima appears. The equivalent water film thickness can becalculated from the following equation (33):

heq ¼

2�n1

� �

arcsin

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�

1þ ½4Rð1��Þ=ð1� RÞ2�

s !

R ¼n1 � n2

n1 þ n2

� �2

� ¼I � Imin

Imax � Imin

� �

ð22Þ

where is the wavelength and n1 and n2 are the bulk refractive indices of thecontinuous and the disperse phases respectively (in the case of foam filmsn2¼ 1); Imax and Imin are the intensity values of the interference maximumand minimum, and I is the instantaneous value of the light intensity. Theabove equation gives the equivalent film thickness, heq (i.e., the film thick-ness plus the thickness of the adsorbed layers calculated by assuming aconstant value of the refractive index equal to n1). A better approximationto the true film thickness can be obtained by correcting for the difference in


refractive index between the bulk film and the adsorbed layer. The correctedfilm thickness is (23)

h ¼ heq � 2hhcn2hc � n21n21 � n22

� �

� 2hpgn2pg � n21

n21 � n22

!

ð23Þ

In Eq. (23), hhc and hhg are the thickness of the region occupied by thesurfactant hydrocarbon chain and polar group, respectively. Similarly, nhcand npg are the corresponding refractive indices. The thickness valuesneeded in order to use Eq. (23) can be estimated from the volume of thetwo parts of the molecule together with values of the area per molecule atthe interface obtained from adsorption data (e.g., the surface-tension iso-therm). Finally, the thickness of the core layer (water in the case of foamfilms) can be calculated as

hcore ¼ h� 2ðhhc þ hpgÞ ð24Þ

The apparatus used for studying thin liquid films is schematicallydepicted in Fig. 6. This device, commonly known as a thin-film pressurebalance, allows drainage patterns of single foam, emulsion, or wetting films

Figure 6 Schematic of the main components of a typical thin-film pressure balance.


to be recorded. The film is formed in a specially constructed cell that isplaced on the stage of an inverted microscope. The reflected light fromthe film is split into two parts: one directed to a charge-coupled device(CCD) camera and another to a fiber-optic probe tip located in the micro-scope eyepiece. The radius of the tip is only about 20 mm, which allows lightfrom a small portion of the film to be registered. The light signal is thenpassed through a monochromatic filter and, finally, directed on to a high-sensitivity photomultiplier. The output of the photomultiplier is connectedto a chart recorder and the data are collected in the form of intensity (as aphotocurrent) versus time. This graph is called an interferogram.

An essential part of the thin-film balance is the cell holding the thinfilm. The cell can be constructed in several ways depending on the type ofmeasurement to be done and the systems under investigation. For, emulsionfilms, the type of cell proposed by Scheludko (33) is often used. The cell isillustrated in Fig. 7. The film is formed between the tips of the menisci of abiconcave drop held in a horizontal tube with radius 1.5–2mm. The tubeand the spiral capillary are filled with the aqueous phase and immersed in acuvette (the lower part of the cell) containing the oil phase. A small suctionpressure applied through the capillary controls the film radius. Recently, acell that is similar to that of Scheludko, but miniaturized about 10-fold wasused by Velev et al. (34). This allowed film sizes and capillary pressuresfound in real emulsion systems to be studied. Bergeron and Radke (35)

Figure 7 Illustration of the Scheludko cell used for investigation of single,

horizontal foam and emulsion films.


used a cell with a porous frit holder as suggested by Mysels and Jones (36)for measuring equilibrium forces across foam and pseudoemulsion films.Their construction is shown in Fig. 8. The main advantage of this so-calledporous-plate technique is that one can vary the pressure in the film bysimply altering the gas pressure in the cell, and thus the stable part ofthe equilibrium disjoining pressure isotherm (where d�=dD < 0) can beobtained.

V. RESULTS AND DISCUSSION

A. Ionic Surfactants on Hydrophobic Surfaces

Many oil-in-water emulsions are stabilized by an adsorbed layer of surfac-tants. One example is asphalt oil-in-water emulsions that often are stabilizedby cationic surfactants (37). The surfactants fulfill two purposes. First, theygenerate long-range repulsive forces which prevent the emulsion droplets

Figure 8 Modified porous-plate cell for investigation of pseudoemulsion films.

(Ref. 35, with permission.)


from coming close to each other. Second, the surfactant layer acts as abarrier against coalescence if the emulsion droplets by chance come closeto each other despite the long-range repulsive forces. The coalescence ishindered by a high spontaneous monolayer curvature, monolayer cohesiveenergy, surface elasticity, and surface viscosity, which increase the activationenergy for hole formation and slow down the depletion of surfactants fromthe contact region. The importance of the cohesive energy for foam filmsstabilized by a homologous series of cationic surfactants was particularlyclearly demonstrated by Bergeron (38). We note that an increased cohesiveenergy in the monolayer increases the bending modulus and thus the freeenergy cost for the surfactant film to have a curvature different than thespontaneous curvature.

Surface force measurements using a hydrophobic solid surface as amodel for a fluid hydrocarbon–water interface provide a good picture of thelong-range forces acting between emulsion droplets. However, the coales-cence behavior of emulsions will not be accurately described from suchmeasurements. One reason is that the fluid interface is much more proneto deformation than the solid surface (facilitating hole formation), and thesurfactant chains can readily penetrate into the fluid oil phase but not intothe solid hydrocarbon surface. Further, the mobility of the surfactants on asolid hydrophobic surface will be different than the mobility at a fluidinterface.

The forces acting between two hydrophobic surfaces across dodecyl-ammonium chloride surfactant solutions are illustrated in Fig. 9 (39). Thelong-range repulsion is due to the presence of an electrostatic double-layerforce. This force is generated by the cationic surfactants that adsorb to thehydrophobic surface thereby generating a surface charge density. The rangeof the double-layer force decreases with increasing surfactant concentration,which is simply a result of the increased ionic strength of the aqueous media.On the other hand, the magnitude of the double-layer force at short separa-tions increases with increasing surfactant concentrations. This is a conse-quence of the increased adsorption of the ionic surfactant that results in anincrease in surface charge density and surface potential. The surfactantconcentration will influence the long-range interactions between oil-in-water emulsions in the same way as observed for the model solid hydro-phobic surfaces; that is, the range of the double-layer force will decrease andthe magnitude of the force at short separations will increase. However, theadsorbed amount at a given surfactant concentration may not be the sameon the emulsion surface as on the solid hydrophobic surface.

At low surfactant concntrations, it is observed that an attraction dom-inates at short separations. The attraction becomes important at separationsbelow about 12 nm when the surfactant concentration is 0.01mM, and


below about 6 nm when the concentration is increased to 0.1mM. Once theforce barrier has been overcome, the surfaces are pulled into direct contactbetween the hydrophobic surfaces at D¼ 0, demonstrating that thesurfactants leave the gap between the surfaces. The solid surfaces havebeen flocculated. However, at higher surfactant concentrations (1mM) thesurfactants remain on the surfaces even when the separation between thesurfaces is small. The force is now purely repulsive and the surfaces areprevented from flocculating. Emulsion droplets interacting in the sameway would coalesce at low surfactant concentrations once they have comeclose enough to overcome the repulsive barrier, but remain stable at highersurfactant concentrations. Note, however, that the surfactant concentrationneeded to prevent coalescence of emulsion droplets cannot be accuratelydetermined from surface force measurements using solid surfaces.

For application purposes, it is often found that asphalt emulsionsstabilized by cationic surfactants function better than such emulsions stabi-lized by anionic surfactants. One main reason is that the interaction betweenthe emulsion droplet and the road material differs depending on the emul-sifier used (37). When the asphalt emulsion is spread on the road surface, itshould rapidly break and form a homogeneous layer. The stones on the roadsurface are often negatively charged and there will be an electrostatic

Figure 9 Force normalized by radius measured between two hydrophobized mica

surfaces in crossed cylinder geometry across aqueous solutions of dodecylammonium

chloride; the surfactant concentration was 0.01mM (g), 0.1mM (f), and 1mM

(�), respectively. The arrows indicate inward jumps occurring when the force barrier

has been overcome. (Redrawn from Ref. 39, with permission.)


attraction between cationic emulsion droplets and the stones. This attrac-tion facilitates the attachment and spreading of the emulsion. On the otherhand, when the emulsion droplet is negatively charged, there will be anelectrostatic repulsion between the stones and the emulsion droplets.

B. Nonionic Surfactants on Hydrophobic Surfaces

Nonionic ethylene oxide-based surfactants are commonly used as emulsi-fiers. Because these surfactants are uncharged, they are not able to generatestabilizing long-range electrostatic forces. Instead, they generate short-rangehydration/protrusion forces that prevent the emulsion droplets from cominginto direct contact with each other. The short-range forces acting betweenhydrophobic solid surfaces coated with such surfactants as a function oftemperature are illustrated in Fig. 10 (40). The zero distance in the diagramis defined as the position of the hard wall (at a value of F/R� 100mN/m).

Figure 10 Force normalized by radius measured between hydrophobized mica

surfaces in crossed cylinder geometry across a 6� 10�5M aqueous solution of

penta(oxyethylene) dodecyl ether. The temperatures were 15�C (g), 20�C (m), 30�C

(^), and 37�C (f). The lines are guides for the eye. (Redrawn from Ref. 40, with

permission.)


The force present at distances above 4 nm is a weak double-layer force. Itoriginates from remaining charges on the hydrophobic substrate surface.The force observed at smaller separations has a pronounced temperaturedependence. It becomes less repulsive with increasing temperature. At thesame time, the adsorbed layer thickness increases, demonstrating that therepulsion between the adsorbed molecules within one layer is also reduced athigher temperatures, facilitating an increased adsorption. The increase inlayer thickness is not seen in Fig. 10 due to our definition of zero separation.

A decreasing interlayer and intralayer repulsion with increasing tem-perature is common for all surfactants and polymers containing oligo(ethyl-ene oxide) groups. This shows that the interaction between ethylene oxidegroups and water becomes less favorable at higher temperatures (i.e., theethylene oxide chain becomes more hydrophobic). There are several theore-tical attempts to explain this phenomenon, but it is outside the scope of thischapter to discuss them and the reader is referred to the original literature(41–48). The temperature dependence of the interaction between oligo(ethyl-ene oxide) chains and water has several important consequences. Themicellar size increases with temperature (49), and the micellar solution hasa lower consolute temperature (i.e., a phase separation occurs onheating) (50).The cloud points for a range of micellar alkyl ethoxylate solutions are

Figure 11 Cloud point temperature of micellar solutions as a function of the

ethylene oxide chain length: the hydrophobic part is an alkyl chain with 8 (g), 10

(^), 12 (m), or 16 (f) carbon atoms. (Data from Ref. 51.) The symbols (œ)

represent the phase-inversion temperature for a 1 : 1 cyclohexane–water emulsion

containing 5% commercial ethylene oxide-based emulsifiers having dodecylalkyl

chains as a hydrophobic group. (Data from Ref. 54.)


provided in Fig. 11 (51). The cloud point increases with the number ofethylene oxide units. The reason is that a longer ethylene oxide chaingives rise to a longer-range intermicellar repulsion and a larger optimalarea per head group (favoring smaller micelles). On the other hand, thecloud point decreases with the length of the hydrocarbon chain. By consid-ering the geometry of the surfactant as described by the packing parameter(52), one realizes that the micellar size is expected to increase withthe hydrocarbon chain length. It is also found that surfaces coated with(ethylene oxide containing) polymers often have good protein-repellentproperties at low temperatures, whereas proteins adsorb more readily tosuch surfaces at higher temperatures (53).

For emulsions the most important aspect may be that the optimal areaper head group in an adsorbed layer decreases with increasing temperature,which reduces the spontaneous monolayer curvature (4). This is the reasonwhy emulsions stabilized by ethylene oxide-based surfactants may changefrom oil-in-water to water-in-oil when the temperature is increased. Thetemperature when this occurs is known as the phase-inversion temperature(PIT). The PIT depends on the length of the hydrocarbon chain and theethylene oxide chain in a manner similar to the cloud point (54) (see Fig. 11).However, the PIT also depends on the type of oil used (55), which is partlydue to differences in solubility of the ethylene oxide surfactants in the dif-ferent oils and to differences in oil penetration in the surfactant layer. Wealso note that if the emulsifier concentration is high enough, a liquid-crystal-line phase may accumulate at the oil–water interface. In such cases, emul-sions which are very stable toward coalescence may be formed (56). This is aresult of the decreased probability of hole formation (4). In this case, thetype of oil used has a dramatic effect on the emulsion stability, which can beunderstood by considering the three-component phase diagram.

C. Nonionic Polymers on Hydrophobic Surfaces

Earlier, we discussed how the length of the oligo(ethylene oxide) chaininfluences the properties of emulsions stabilized by alkyl ethoxylates.When the ethylene oxide chain becomes sufficiently long, one normallyrefers to the substance as a diblock copolymer rather than as a surfactant.Of course, there is no clear distinction, but the properties vary in a con-tinuous fashion with increasing ethylene oxide chain length. It is of interestto follow how the forces acting between two surfaces carrying adsorbeddiblock copolymers vary with the length of the adsorbing (anchor) blockand the nonadsorbing (buoy) block. A nice experimental work addressingthis question is that of Belder et al (57). Fleer et al. give a thorough theore-tical treatment in their book (17), where it was suggested that the most


efficient steric stabilization is obtained when the anchor block has a size thatis 10–20% of that of the buoy block. The reason for this optimum is thatwhen the anchor block is too small, the driving force for adsorption is weakand the adsorbed amount will be low. On the other hand, when the anchorblock is too large, the area per adsorbed molecule will be large. As a con-sequence, the buoy block layer will be dilute and it will not extend very farout into the solution, leading to a not very pronounced steric force.

The forces acting between solid hydrophobic surfaces coated withdifferent ethylene oxide-based diblock polymers are illustrated in Fig. 12.The forces acting between surfaces coated with penta(oxyethylene) dodecylether, C12E5, becomes significantly repulsive at distances below about 2 nm,calculated from the hard wall contact at D¼ 0. Note that the surfactantlayer remains between the surfaces, and the range of the force given is thusrelative to the position of direct contact between the compressed adsorbedsurfactant layers. The forces between hydrophobic surfaces coated with adiblock copolymer containing 8 butylene oxide units and 41 ethylene oxideunits, B8E41, are significantly longer ranged. The interaction at distancesabove 4 nm from the ‘‘hard wall’’ is dominated by a weak electrostaticdouble-layer force originating from remaining charges on the silanatedglass surface. However, at shorter distances, a steric force predominates.

Figure 12 Force normalized by radius between hydrophobized mica or glass

surfaces coated by penta(oxyethylene) dodecyl ether at 20�C (g), and copolymers of

butylene oxide (B) and ethylene oxide (E) with compositions B8E41 (lower line) and

B15E200 (upper line). All data have been recalculated to spherical geometry.


Hence, the molecules with the longer ethylene oxide chains give rise to alonger-range force. Note that this is true even when the range is calculatedfrom the position of the hard wall (i.e., without considering the difference incompressed layer thickness). A much longer-range force is observed in thecase of B15E200, where the steric force extends to more than 10 nm awayfrom the hard wall contact.

From the above, it is clear that rather large diblock copolymers areefficient in generating long-range repulsive steric forces, which is beneficialfor increasing the stability of dispersed particles and emulsion droplets. Aneven higher stability may be obtained if a mixture of diblock copolymersand charged surfactants are used, thus providing both steric and electro-static stabilization.

D. Polyelectrolytes on Surfaces

Both steric and electrostatic stabilization was utilized by Faldt et al. (58)when making soybean oil emulsions. They first made the emulsion using amixture of phosphatidylcholine and glycoholic acid (bile salt) with a pKa of4.4. The emulsion droplets obtained a net negative surface charge due todissociation of the glycoholic acid. To improve the stability of the emulsiona weak cationic polyelectrolyte, chitosan, with a pKa of 6.3–7 was added.The polyelectrolyte adsorbs to the negatively charged emulsion dropletsurface, which becomes positively charged at low pH. It was found thatthe emulsion was stable at high and low pH but not at pH values around7, where irreversible aggregation was observed. This clearly shows thatthe forces acting between the emulsion droplets change with pH. Toshed light on this behavior, the forces acting between negatively chargedsolid surfaces coated by chitosan were measured as a function of pH(Fig. 13).

A repulsive double-layer force dominated the long-range interaction atpH values below 5. However, at distances below about 5 nm, the measuredrepulsive force is stronger than expected from DLVO theory due to thepredominance of a steric force contribution. The layer thickness obtainedunder a high compressive force was 1 nm per surface. Hence, it is clear thatpositively charged chitosan adsorbs in a very flat conformation on stronglyoppositely charged surfaces such as mica with only short loops and tails.When the pH is increased to 6.2, the mica–chitosan system becomesuncharged, because the charge density of the chitosan molecules hasdecreased. The decrease in charge density of the chitosan also results in adecrease in segment–segment repulsion and, therefore, an even more com-pact adsorbed layer. At this pH value, there is an attraction between thelayers at a surface separation of about 2 nm. The steric repulsion is, in this


case, very short range (<2nm) and steep. A further increase in pH to 9.1results in a recharging due to the fact that the charges on the polyelectrolyteno longer can compensate for all of the mica surface charges. Further, as thecharge density of the polyelectrolyte is reduced, the range of the steric forceincreases again due to the lower affinity of the polyelectrolytes for thesurface. Clearly, the mica–chitosan system is positively charged at low pH(i.e., the charges on the polyelectrolyte overcompensate for the charges onthe mica surface), uncharged at pH 6.2, and negatively charged at highpH due to an undercompensation of the mica surface charge by thepolyelectrolyte charges.

The flocculation behavior of the soybean emulsion can now be betterunderstood. A stable emulsion is formed by a low pH owing to the electro-static repulsion generated by the excess charges from the adsorbed chitosan.An intermediate pH values, the soybean emulsion is uncharged and theadsorbed chitosan layer is very flat. Hence, no long-range electrostaticforce or long-range steric force is present that can stabilize the emulsion.At a high pH, the charges due to ionization of the glycoholic acid are nolonger compensated for by the high pH at nearly uncharged chitosan. Thus,stablizing electrostatic forces are once again present. Further, the range ofthe stabilizing steric force is most likely also increased.

Figure 13 Force normalized by radius between negatively charged mica surfaces in

crossed cylinder geometry precoated with a layer of chitosan, a cationic poly-

electrolyte. The forces were recorded at pH 3.8 (^), 4.9 (^), 6.2 (e), 7.6 (œ), and

9.1 (g); the arrow indicates an outward jump. (From Ref. 58, with permission.)


E. Proteins on Hydrophobic Surfaces

Amphiphilic proteins have properties similar to those of block copolymersand surfactants in the sense that they have clearly separated hydrophilicand hydrophobic domains that allow the formation of monodisperseaggregates or micelles in solution. For b-casein, the association processstarts at a protein concentration of around 0.5mg/mL at room tempera-ture (59). Amphiphilic proteins adsorb strongly to nonpolar surfaces incontact with aqueous solutions, and they may generate stabilizing stericand electrostatic forces. In fact, caseins isolated from milk are widely usedin different technical products ranging from food to paint and glue. Onereason for this is that the caseins have excellent properties as emulsifiersand foaming agents, and emulsions stabilized by proteins constitute themost important class of food colloids. The caseins protect the oil dropletsfrom coalescing and also provide long-term stability during storage andsubsequent processing (60). b-Casein is more hydrophobic compared to theother caseins and the charged domain is clearly separated from thehydrophobic part, which makes the b-casein molecule, as whole, distinctlyamphiphilic (61). At pH 7, the isolated b-casein molecule carries a netcharge of about �12 (61).

Nylander and co-workers investigated the interactions betweenadsorbed layers of b-casein in order to clarify the mechanism responsiblefor the ability of b-casein to act as a protective colloid (62,63). The force as afunction of surface separation between hydrophobic surfaces across a solu-tion containing 0.1mg/mL b-casein and 1mM NaCl (pH¼ 7) is illustratedin Fig. 14. At separations down to about 25 nm, an electrostatic double-layer force dominates the interacation. The hydrophobic substrate surfacewas uncharged, so the charges responsible for this force had to come fromthe adsorbed protein. When the surfaces are compressed closer together, therepulsive force is overcome by an attraction at a separation of about 25 nm,and the protein-coated surfaces are sliding into contact about 8 nm out fromthe hydrophobized mica surface (Fig. 14, inset). This observation, as well asthe adhesive force found on separation, was interpreted as being due tobridging of the hydrophilic tails that extend out into solution. Furthercompression does not significantly change the surfaces separation. Theresults indicate that the adsorbed b-casein layer consists of an inner compactpart and a dilute outer region.

This conclusion compares favorably to what is known from studies ofthe adsorption of b-casein on to air–liquid, liquid–liquid, and solid–liquidinterfaces using a range of other techniques. It has generally been found thatthe adsorbed amount of b-casein on hydrophobic surfaces is between 2 and3mg/m2 over a wide range of bulk concentrations. This is the case for planar


air–water and planar oil–water interfaces (59), for hydrocarbon oil–waterinterfaces in emulsions (64), and for interfaces between water and poly-styrene latex particles (65–67) and hydrophobized silica (68). At thetriglyceride–water interface, however, the adsorbed amount is somewhatlower (69).

Information about the adsorbed layer structure of b-casein at thehydrophobic surface can be obtained by employing neutron reflectivity,small-angle x-ray scattering (SAXS), and dynamic light scattering. It wasfound that the layer of b-casein adsorbed to a hydrocarbon oil–waterinterface or an air–water interface (70,71) consisted of a dense inner part,2 nm thick, and a protein volume fraction of 0.96, immediately adjacent tothe interface. Beyond that, a second dilute region with a protein volumefraction of 0.15 extended into the aqueous phase. A similar structure of b-casein adsorbed onto polystyrene latex particles was observed with SAXS(65). The electron density profile calculated from the SAXS data indicatedthat most of the protein resided within 2 nm from the surface. The profile

Figure 14 Normalized force measured between hydrophobized mica surfaces in

crossed cylinder geometry coated with b-casein in a solution containing 0.1mg

b-casein/mL (pH¼ 7; 1mM NaCl) (^,^) and after dilution with 1mM NaCl

(f,e). Filled and unfilled symbols represent the force measured on compression

and decompression, respectively; m represents, the force measured between hydro-

phobized mica surfaces across a 0.1-mM NaCl solution at pH 5.6 containing 0.2mg

proteoheparan sulfate/mL. The inset shows the measured forces between adsorbed

layers of b-casein before and after dilution with 1mM NaCl on an expanded scale.

(From Ref. 26, with permission.)


also showed a small amount of protein extending some 10 nm into theaqueous phase. Further, the hydrodynamic layer thickness estimated fromthe diffusion coefficient determined by dynamic light scattering of latexparticles (67,72,73) and emulsion droplets (69) coated with b-casein wasfound to be 10–15 nm. The fact that different experimental techniquesgive a different value of the layer thickness is simply because they have adifferent sensitivity to the extending tails.

This type of layer structure, with a compact inner region and a diluteouter region, was also predicted by self-consistent field theory and bycomputer simulations. For instance, Monte Carlo simulations show that adense layer (1–2 nm thick) is present close to the planar interface (74). Thislayer contained about 70% of the segments. Further out, a region of muchlower density was found to extend about 10 nm into the aqueous phase.Similar results were obtained by self-consistent field calculations (75),which also showed that the most hydrophilic segments reside predominantlyin the outer layer.

The properties of adsorbed b-casein layers can be changed by theaction of enzymes. Leaver and Dalgleish (69) have observed that the N-terminal end is more accessible to trypsinolysis than the rest of the adsorbedmolecule, and that loss of the tail leads to a reduced layer thickness. Asimilar change was observed by Nylander and Wahlgren (68), who foundthat the addition of endoproteinase ASP-N to an adsorbed layer of b-caseinreduced the adsorbed amount by approximately 20%. The removal ofthe extending tails will clearly reduce the range of the stabilizing stericforce and thus reduce the emulsion stability against flocculation. Wenote that the forces generated by adsorbed b-casein are not very stronglyrepulsive (Fig. 14). Hence, the excellent stability of emulsion dropletscoated by b-casein is most likely because the hole nucleation energy barrieris high.

Proteoheparan sulfate is an amphiphilic membrane glycoprotein.Like b-casein, it has one large hydrophobic region. Proteoheparan sulfatehas three to four hydrophilic and strongly charged side chains, whereasb-casein has only one less charged tail. Protoheparan sulfate is not usedfor stabilizing emulsions. However, it is nevertheless of interest to comparethe forces generated by this protein with those generated by b-casein.The interaction between hydrophobic surfaces across a 1-mM NaCl solu-tion containing 0.2mg proteoheparan sulfate/mL is shown in Fig. 14(62,76). The long-range interaction is dominated by a repulsive double-layer force, considerably stronger than that observed for b-casein. Thisis simply because proteoheparan sulfate is more strongly charged thanb-casein. A steric force dominates the short-range interaction for bothproteins.


F. Phospholipids on Polar Surfaces in Oil

We have seen earlier that surface force measurements provide importantinformation about interactions between solid hydrophobic surfaces coatedwith surfactants and polymers, and that some of the information obtained isdirectly relevant for oil-in-water emulsions. However, the details of theinteraction profiles are expected to be different for liquid hydrocarbondroplets coated with the same molecules as the model solid surfaces. Inparticular, the coalescence behavior of the emulsion droplets cannot bemodeled. It is even more difficult to make a solid model surface thatmimicks the behavior of water-in-oil emulsions. At present, the best onecan do is to use a polar surface that attracts the polar part of the emulsifier.In this way, the orientation of the emulsifier on the model surface and at thewater-in-oil emulsion surface will be the same. This will allow us to drawsome conclusions about how polar solid surfaces coated with emulsifiersinteract across oil, but care should be taken when using such results todraw conclusions about water-in-oil emulsions.

The forces between polar mica surfaces interacting across triolein con-taining 200 ppm of soybean phosphatidylethanolamine (PE) have beenstudied (77). Some results obtained at two different water activities areillustrated in Fig. 15. When the water activity is 0.47, a monolayer of PEis adsorbed on each surface. The orientation is such that the polar group isattached to the mica surface with the nonpolar part directed toward the oilphase. Thus, adsorption of the phospholipid renders the mica surface non-polar. No force is observed until the surfaces are about 6 nm apart. Atthis point, a very steep repulsion is experienced which is due to compressionof the adsorbed layers. A weak attraction is measured on separation.The forces change significantly when the trioein is saturated with water(water activity¼ 1). The adsorbed layer becomes significantly thinner, andnow only a rather weak compressive force is needed in order to merge thetwo adsorbed layers into one. The reason is that water molecules adsorbnext to the polar mica surface and in the region of the zwitteronic lipidhead group. This increases the mobility of the adsorbed phospholipid anddecreases the force needed to merge the two adsorbed layers. Interestingly, itis not possible to remove the last adsorbed layer even by employing a veryhigh compressive force.

From these observations, we can draw some conclusions that are rele-vant for water-in-oil emulsions. First, no long-range electrostatic forces arepresent in the nonpolar media. This is because the dissociation of surfacegroups is very unfavorable in low-polarity media. Hence, generally it is verydifficult to utilize electrostatic forces for generating long-range stabilizingforces in oil. Surfactants like phospholipids or alkyl ethoxylates adsorbed


in monolayers will only generate short-range repulsive forces due tocompression of the hydrocarbon chains penetrating into the oil medium.These substances will be efficient in preventing coalescence of water-in-oil emulsions only when the adsorbed amount is high enough and thespontaneous monolayer curvature is sufficiently negative.

G. Polymers on Polar Surfaces in Oil

We saw earlier that surfactants adsorbed in monolayers only give rise torather short-range forces in oil media. The range of the forces can beincreased considerably if liquid-crystalline phases are accumulated at theinterface, or if amphiphilic oil-soluble polymers are used instead of low-molecular-weight surfactants. An example of such a polymer is PGPR(polyglycerol polyricinoleate), which is a powerful water-in-oil emulsifierused in the food industry (78). PGPR has a complex branched structureas indicated in Fig. 16. This polymer was used for studying interactionsbetween polar mica surfaces in triolein (79). The forces obtained at a poly-mer concentration of 200 ppm are shown in Fig. 17. In this system, repulsivesteric forces are observed at distances below 15 nm. The magnitude of the

Figure 15 Force normalized by radius between mica surfaces in crossed cylinder

geometry interacting across a triolein solution containing 200 ppm of soybean

phosphatidylethanolamine. The forces of soybean were measured at water activities

of 0.47 on approach (g) and separation (œ), as well as at a water activity of 1 on

approach (f) and separation (e); the arrows indicate inward and outward jumps.

(From Ref. 77, with permission.)


force increases rather slowly with decreasing surface separation until thesurfaces are about 5 nm apart. A further compression of the layers resultsin a steep increase of the steric force. The force profile indicates that theadsorbed layer consists of an inner dense region and an outer diluteregion with some extended tails and loops. When dense polymer layersthat generate long-range steric forces and have a high surface elasticityand viscosity are adsorbed at the interface of water-in-oil emulsions, onecan expect that the emulsion stability against flocculation and coalescencewill be good.

Figure 17 Force normalized by radius between mica surfaces interacting across a

triolein solution containing 200 ppm of PGPR measured on approach.

Figure 16 Illustration of the structural elements of PGPR. The upper structure is

that of the polyricinoleate moiety; the lower structure shows the polyglycerol

backbone. The R in the structure can be either hydrogen, a fatty acid residue, or a

polyricinoleate residue. In PGPR, at least one of the side chains is polyricinoleate.


H. Forces Between Surfaces Across Emulsions

Emulsion droplets do not only break by coalescing with each other, but theymay also break by attaching to a solid surface. Depending on the applica-tion, this may be wanted or unwanted. In order to study emulsion–surfaceinteractions, a model oil-in-water emulsion was prepared from purified soy-bean oil (20wt%) using fractionated egg phosphatides (1.2 wt%) as theemulsifier. The mayor components of the emulsifier were phosphatidylcho-lines and PEs. The mean diameter (DZ average) of the emulsion was 320 nm,as determined with photon-correlation spectroscopy. A small amount ofnegatively charged lipids was also present, giving the emulsion droplets anet negative zeta potential of about �40mV (80). This emulsion was thenplaced inside a SFA.

The forces acting between two glass surfaces across the 20% oil-in-water emulsion measured by using the MASIF are illustrated in Fig. 18 (81).A repulsive force dominates the interactions at separations below 200 nm.

Figure 18 Force normalized by local geometric mean radius as a function of

surface seperation between glass across a concentrated emulsion solution (20wt% oil

and 1.2wt% phospholipid). The thinner lines correspond to the force measured on

separation, the dashed line represents the calculated force between two spherical

surfaces connected by a capillary condensate in the full equillibrium case [Eq. (25)],

and the dotted line represents the force between two spherical surfaces connected by

a capillary condensate in the nonequillibrium case [Eq. (26)]. (From Ref. 81, with

permission.)


The force increases strongly with decreasing distance. This illustrates thatlarge aggregates, with a diameter of at least 100 nm, are associated witheach surface and the repulsion between 40 and 200 nm is due to deforma-tion and eventual breaking of these aggregates. The range of the repulsiveforce is consistent with the layer thickness obtained in a previous ellipso-metric study by Malmsten et al. (80). They found that the thickness of alayer adsorbed from the emulsion on to a negatively charged silica surfacewas around 100 nm, independent of surface coverage.

In some force curves, one or two distinct steps are present. Figure 18illustrates one such force curve where a clear step is seen at a separation ofabout 40 nm. At a separation of about 10 nm, another step, but less pro-nounced, is seen. These steps are interpreted as being due to coalescence ofadsorbed emulsion droplets and/or due to materials that collectively leavethe zone between the surfaces. On subsequent approaches of the surfaces onthe same position, the range of the repulsion remains at about 200 nm.However, the steps in the force profile become less pronounced or disappearcompletely, indicating a change in the adsorbed layer when exposed to ahigh compressive force.

A strong and long-range force is observed when the surfaces are sepa-rated. It is plausible that this attraction is due to the formation of a capillarycondensate of oil between the surfaces (Fig. 19). This capillary condensateoriginates from the emulsion droplets that have been destroyed when thesurfaces are brought together. The forces between two spherical surfacesconnected by a capillary condensate in the full equilibrium case are givenby (82)

F

R¼ 2�ðsc � sbÞ 1�

D

Rk

� �

ð25Þ

Figure 19 Schematic of the capillary condensate formed between glass surfaces

due to breakdown of adsorbed emulsion droplets. The figure is not according to

scale. (From Ref. 81, with permission.)


where is the interfacial tension and the subscript s, c, and b stand forsurface, capillary condensate, and bulk, respectively; RK is the Kelvin radiusof the capillary condensate. In cases when the surfaces are separated toorapidly to allow the volume of the capillary to change with separation, oneinstead obtains (82)

F

R¼ 2�ðsc � sbÞ 1�

1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffið1þ R2

k=D2Þ

q

2

64

3

75 ð26Þ

Two theoretical force curves calculated by using Eqs. (25) and (26) areshown in Fig. 18. In these calculations, we used a Kelvin radius of 320 nmand an interfacial tension difference of 3.3mN/m. The measured forcecurves fall in between the extreme cases of full equilibrium, where thevolume of the condensate is changing with distance to minimize the freeenergy, and the case of no change in condensate volume with separation.Long-range forces due to capillary condensation have been observedpreviously by Petrov et al., who found that a lamellar phase condensedbetween two surfaces immersed in an L3-phase (83). Capillary condensationof sparingly soluble surfactants between surfaces close to each other insurfactant solutions has also been reported (84).

It is worth pointing out that the functional form of the measuredattraction shows that the volume of the capillary condensate decreaseswith increasing separation. However, this does not occur fast enough com-pared to the speed of the measurements (the attractive part took about 30 sto measure) to allow full equilibrium to be established. Also, the range of themeasured repulsion on approach does not increase with the number oftimes the surfaces are brought into contact but rather the reverse. Bothof these observations point to the fact that the material present in thecapillary condensate is spontaneously reemulsified when the surfaces areseparated.

In order to obtain information about whether a monolayer, a bilayer,or a multilayer was firmly attached to the surfaces, we employed the inter-ferometric SFA and mica surfaces rather than glass surfaces (81). In thesemeasurements, a drop of the emulsion was placed between the surfaces. Theemulsion was very opaque and no interference fringes could be seen until thesurfaces were close to contact. The force measured between mica surfacesacross a concentrated emulsion were repulsive and long range (several hun-dred nanometers), which was in agreement with the results obtained usingglass surfaces. Because the forces were so highly repulsive, no attempt wasmade to measure them accurately, but under a high compressive force, the


surfaces come to a separation 8.5 nm. This corresponded to a bilayer ofphospholipid on each surface.

I. Forces Due to Stratification in Foam and Pseudoemulsion Films

The thinning of thin liquid films in micellar solution is found to occur in astepwise fashion, known as stratification. Bergeron and Radke (35) set out tostudy the forces responsible for this phenomenon using the porous frit ver-sion of the thin-film pressure balance. They found that the equilibrium dis-joining pressure curve (force curve) showed an oscillatory behavior both forfoam and pseudoemulsion (i.e., asymmetric oil–water–gas) films stabilizedby the anionic surfactant sodium dodecyl sulfate above the cmc (Fig. 20).The reason for this oscillatory force profile was the layering of micelles in theconfined space in the thin aqueous film separating the two interfaces. Theperiodicity of the oscillations was the same for foam films and for pseudoe-mulsion films. The main difference between the two systems was found in thehigh-pressure region of the disjoining pressure isotherm. The pseudo-emulsion films ruptured at much lower imposed pressures than the foamfilms. This was attributed to the action of the oil phase as a foam destabilizer.

Figure 20 Low-pressure region of the disjoining pressure isotherm across a 0.1M

SDS solution in a single-foam lamella, and across a 0.1M SDS solution separating a

dodecane–solution interface from an air–solution interface. (Reproduced from

Ref. 35, with permission.)


VI. SUMMARY

Several techniques are available for studying long-range interactionsbetween solid surfaces and fluid interfaces. The forces generated by surfac-tants, polymers, and proteins have been determined. For oil-in-water emul-sions, both steric and electrostatic stabilizing forces are of importance,whereas only steric forces are operative for the case of water-in-oil emul-sions. These forces are well understood theoretically. The experimental tech-niques employed give very detailed information on the long-range forces,and in this respect, the results obtained for the model systems can be usefulfor understanding interactions in emulsion systems. However, the surfaceforce techniques employed are not suitable for modeling the molecularevents leading to coalescence of emulsion droplets once they have beenbrought in close proximity to each other. Some data illustrating the break-down and reemulsification of emulsion droplets in the gap between twomacroscopic solid surfaces were also presented. This is a new researchtopic and very little is known about how the surface properties and thetype of emulsifier influence the stability of emulsion droplets at surfacesand in narrow gaps between surfaces.

ACKNOWLEDGMENT

This work was partly sponsored by the Competence Centre for SurfactantsBased on Natural Products (SNAP).

REFERENCES

1. W. D. Bancroft, J. Phys. Chem. 19, 275 (1915).

2. W. C. Griffin, J. Soc. Cosmet. Chem 1, 311 (1949).

3. J. T. Davies, Proc. 2nd Int. Congr. Surf. Activity 1957, Vol. 1, 426.

4. A. Kabalnov and H.Wennerstrom, Langmuir 12, 276 (1996).

5. D. Y. C. Chan and R. G. Horn, J. Chem, Phys. 83, 5311 (1985).

6. J. Kizling and B. Kronberg, Colloids Surfaces 50, 131 (1990).

7. O. Reynolds, Phi. Trans. R. Soc. London 177, 157 (1886).

8. F. W. Cain and J. C. Lee, J. Colloid Interf. Sci. 106, 70 (1985).

9. E. D. Manev and R. Tsekov, and B. P. Radoev, J. Dispers. Sci. Technol. 18, 769

(1997).

10. B. P. Radoev, A. D. Scheludko, and E. D. Manev, J. Colloid. Interf. Sci. 95, 254

(1983).

11. S. A. K. Jeelani and S. Hartland, J. Colloid Interf. Sci. 164, 296 (1994).


12. J. N. Israelachvilli, Intermolecular and Surface Forces, Academic Press,

London, 1991.

13. E. J. N. Verwey and J. T. G. Overbeek, Theory of Stability of Lyophobic

Colloids, Elsevier, Amsterdam, 1948.

14. B. V. Derjaguin and L. Landau, Acta Physicochim. URSS 14, 633 (1941).

15. V. A. Parsegian, N. Fuller, and R. P. Rand, Proc. Natl. Acad. Sci. USA 76,

2750 (1979).

16. J. N. Israelachvili and H. Wennerstrom, J. Phys. Chem. 96, 520 (1992).

17. G. J. Fleer, M. A. Cohen Stuart, J. M. H. M. Scheutjens, T. Cosgrove, and

B. Vincent, Polymers at Interfaces, Chapman & Hall, London, 1993.

18. P.-G. de Gennes, Scaling Concepts in Polymer Physics, Cornell University Press,

Ithaca, NY, 1979.

19. P.-G. de Gennes, Adv. Colloid Interf. Sci. 27, 189, (1987).

20. A. D. Nikolov and D. T. Wasan, J. Colloid Interf. Sci. 133, 1–12, (1989).

21. V. Bergeron and C. J. Radke, Langmuir 8, 3020 (1992).

22. V. Bergeron, A. J. Jimenez-Laguna, and C. J. Radke, Langmuir 8, 3027 (1992).

23. V. Bergeron, Forces and structure in surfactant-laden thin-liquid films, PhD

thesis, University of California, Berkeley, 1993.

24. A. L. de Vries, Rec. Trav. Chim. Pays-Bas 77, 383 (1958).

25. K. Shinoda and H. Saito, J. Colloid Interf. Sci. 30, 258 (1969).

26. P. M. Claesson, T. Ederth, V. Bergeron, and M. W. Rutland, Adv. Colloid

Interf. Sci.67, 119 (1996).

27. J. N. Israelachvili, G. E. Adams, J. Chem. Soc. Faradav Trans. I 74, 975 (1978).

28. J. Parker, Prog. Surf. Sci. 47, 205 (1994).

29. K. Schillen, P. M. Claesson, M. Malmsten, P. Linse, and C. Booth, J. Phys.

Chem. 101, 4238 (1997).

30. B. Derjaguin, Kolloid Z. 69, 1557 (1934).

31. R. G. Horn, J. N. Israelachvili, and F. Pribac, J. Colloid Interf. Sci. 115, 480

(1987).

32. P. Attard and I. L. Parker, Phys. Rev. A 46, 7959 (1992).

33. A. Scheludko, Adv. Colloid Interf. Sci. 1, 391 (1967).

34. O. D. Velev, G. N. Constantinides, D. G. Avraam, A. C. Paytakes, and R. P.

Borwankar, J. Colloid Interf. Sci. 175, 68 (1995).

35. V. Bergeron and C. J. Radke, Colloid Polym. Sci. 273, 165 (1995).

36. K. J. Mysels and M. N. Jones, Disc. Faraday Soc. 42, 42 (1966).

37. R. L. Ferm, in Emulsion and Emulsion Technology (K. J. Lissant, ed.), Marcel

Dekker, New York, 1974, Vol I p. 387.

38. V. Bergeron, Langmuir 13, 3474 (1997).

39. P. C. Herder, J. Colloid Interf. Sci. 134, 336 (1990).

40. P.M. Claesson, R. Kjellander, P. Stenius, and H. K. Christenson, J. Chem. Soc.

Faraday Trans. I 82, 2735 (1986).

41. R. Kjellander, J. Chem. Soc. Faraday Trans. II 78, 2025 (1982).

42. R. E. Goldstein, J. Chem. Phys. 80, 5340 (1984).

43. G. Karlstrom, J. Phys. Chem. 89, 4962 (1985).

44. A. Matsuyama and F. Tanaka, Phys. Rev. Lett. 65, 341 (1990).


45. P.-G. de Gennes, C.R. Acad. Sci. Paris 313, 1117 (1991).

46. S. Bekiranov, R. Bruinsma, and P. Pincus, Europhys. Lett. 24, 183 (1993).

47. P. Linse and B. Bjorling, Micromolecules 24, 6700 (1991).

48. M. Bjorling, Macromolecules 25, 3956 (1992).

49. W.Brown,R. Johnsen, P. Stilbs, andB.Lindman, J. Phys. Chem. 87, 4548 (1983).

50. G. J. T. Tiddy, Phys. Rep. 57, 3 (1980).

51. D. J. Mitchell, G. J. T. Tiddy, L. Waring, T. Bostock, and M. P. McDonald,

J. Chem. Soc. Faraday Trans. I 79, 975 (1983).

52. J. N. Israelachvili, D. J. Mitchell and B. W. Ninham, J. Chem. Soc. Faraday

Trans. II 72, 1525 (1976).

53. S. I. Jeon and J. D. Andrade, J. Colloid Interf. Sci. 142, 159 (1991).

54. K. Shinoda and H. Takeda, J. Colloid Interf. Sci. 32, 642 (1970).

55. B. A. Bergenst_aahl and P. M. Claesson, in Food Emulsions (S. E. Friberg and

K. Larsson, eds), Marcel Dekker, New York, 1997.

56. S. E. Friberg and C. Solans, Langmuir 2, 121 (1986).

57. G. F. Belder, Gten Brinke and G. Hadziioannou, Langmuir 13, 4102 (1997).

58. P. Faldt, B Bergenst_aahl and P. M. Claesson, Colloids Surfaces A 71, 187 (1993).

59. D. G. Schmidt and T. A. J. Payens, J. Colloid Interf. Sci. 39, 655 (1972).

60. E. Dickinson, J. Dairy Sci. 80, 2607 (1997).

61. H. E. Swaisgood, Development in Dairy chemistry— 1, Applied Science,

London, 1982.

62. P. M. Claesson, E. Blomberg, J. C. Froberg, T. Nylander, and T. Arnebrandt,

Adv. Colloid Interf. Sci. 57, 161 (1995)

63. T. Nylander and N. M. Wahlgren, Langmuir 10, 1274 (1997).

64. J.-L. Courthaudon, E. Dickinson, and D. G. Dalgleish, J. Colloid Interf. Sci.

145, 390 (1991).

65. A. R. Mackie, J. Mingins, and A. N. North, J. Chem. Soc. Faraday Trans. 87,

3043 (1991).

66. J. R. Hunter, P. K. Kilpatrick, and R. G. Carbonell, J. Colloid Interf. Sci. 142,

429 (1991).

67. D. V. Brooksbank, C. M. Davidson, D. S. Horne, and J. Leaver, J. Chem. Soc.

Faraday Trans. 89, 3419 (1993).

68. T. Nylander and N. M. Wahlgren, J. Colloid Interf. Sci. 162, 151 (1994).

69. J. Leaver and D.G. Dalgleish, J. Colloid Interf. Sci. 149, 49 (1992).

70. E. Dickinson, J. Chem. Soc. Faraday Trans. 88, 2973 (1992).

71. E. Dickinson, D. S. Horne, J. S. Phipps, and R. M. Richardson, Langmuir 9,

242 (1993).

72. D. G. Dalgeish, Colloids Surfaces 46, 141 (1990).

73. D. G. Dalgleish and J. Leaver, J. Colloid Interf. Sci.141, 288 (1991).

74. E. Dickinson and S. R. Euston, Adv. Colloid Interf. Sci. 42, 89 (1992).

75. F A. M. Leermakers, P. J. Atkinson, E. Dickinsin, and D. S. Horne, J. Colloid

Interf. Sci. 178, 681 (1996).

76. M. Malmsten, P. M. Claesson, and G. Siegel, Langmuir 10, 1274 (1994).

77. A. Dedinaite, P. M. Claesson, B, Campbell, and H. Mays. Langmuir 14, 5546

(1998).


78. R. Wilson, B. I. van Schie, and D. Howes, Food Chem. Toxicol. 36, 711 (1998).

79. A. Dedinaite, B. Campbell, Langmuir 16, 2248–2253 (2000).

80. M. Malmsten, A.-L. Lindstrom, and T. Warnheim, J. Colloid Interf. Sci. 173,

297 (1995).

81. E. Blomberg, P. M. Claesson, and T. Warnheim, Colloids Surfaces A, 159,

149–157 (1999).

82. D. F. Evans and H. Wennerstrom, The Colloidal Domain, 2nd ed., VCH,

New York, 1998.

83. P. Petrov, U. Olsson, and H. Wennerstrom, Lanfmuir 13, 1000, (1979).

84. A. Waltermo, P. M. Claesson, and I. Johansson, J. Colloid Interf. Sci. 183, 506

(1996).


8Coalescence Mechanisms inProtein-Stabilized Emulsions

George A. van AkenWageningen Centre for Food Sciences, Wageningen, The Netherlands

I. INTRODUCTION

A. Importance of Coalescence in Food Emulsions

Coalescence of emulsified oil may be defined as any physical process thatleads to the merging of the oil droplets into a larger droplet. Coalescence isdesired in some cases—for example, during butter manufacture, wherechurning of cream leads to the separation of butter granules through aphase-inversion process that involves coalescence of milk fat globules (1).Another example is during whipping of cream (2), where coalescence leadsto the formation of a network of partially coalesced fat globules, whichstabilizes the air bubbles in the cream (1,3). A final example is a limitedamount of coalescence in the mouth, which may lead to the release of someoil that may coat the oral tissues and provide oily lubriciy (4) and enhancethe release of fat-soluble flavours (5). However, usually coalescence is unde-sired, such as during emulsion formation by homogenisation, where theoccurrence of coalescence of newly formed droplets reduces the efficiencyof the homogenizer (6), and during storage, where coalescence leads to theformation of larger droplets with an increased creaming rate or even avisible oil layer on top of the emulsion.

B. Adsorbed Protein Layers

In food emulsions, stability against coalescence is obtained from anadsorbed layer at the surface of the oil droplets. The composition of this


layer is usually complex, often consisting primarily of protein and proteinaggregates, in combination with polar lipids, such as lecithin, and a widerange of low-molecular-weight surfactants, such as monoglycerides anddiglycerides, polysorbates, Tweens, Spans, and sucrose esters (7,8). Themost commonly used proteins in food applications are the caseins andwhey proteins derived from milk. Caseins are a group of proline-rich pro-teins with relatively little secondary structure and low solubility at pH 4.6(2,9). In solution, they are often present in the form of micelles that solu-bilize calcium phosphate. Whey proteins are a group of globular proteins(2,9) which have in common a much more compact and ordered molecularstructure, organized in tight conformations, such as a-helices and b-sheets,and stabilized by physical bonds (e.g., hydrogen bonds, hydrophobic bonds)and covalent bonds (mainly disulfide bridges).

After proteins have adsorbed to the oil–water interface of the emulsiondroplets, they slowly change their conformation (‘‘unfold’’), adapting theirmolecular structure to the changed environment at the interface (10–13). Ithas been suggested that on adsorption, globular protein molecules largelyretain their secondary structure of a-helices and b-sheets but that the orga-nization of these elements (tertiary structure) changes. This reorganizaitonleads to an increase of the binding energy of the protein to the interface andexposure of hydrophobic moieties and sulfhydryl groups at the aqueous sideof the adsorbed protein layer (14,15). Intermolecular bonds are formedbetween the adsorbed molecules, largely hydrophobic bonds but, formany proteins, probably also intermolecular disulfide bonds (16). Thisleads to the formation of a cohesive layer of the adsorbed molecules, withstrongly reduced lateral mobility of the molecules and the tendency of theadsorbed layer to fracture rather than flow when the adsorbed layer isdeformed (17).

Some compounds bind to adsorbed protein layers. (e.g., polyphenols(tannins) and many polysaccharides, such as propylene glycol alginate,pectin and carrageenan). This was shown to enhance the elasticity of amixed adsorbed layer of protein and low-molecular-weight surfactantsand to reduce the rate of drainage of foams stabilized by such a mixedadsorbed layer (18).

The presence of coadsorbed surfactants and polar lipids oftenimproves the overall performance of protein- or peptide-stabilized emul-sions. These surfactants help to lower the surface tension during emul-sification. Moreover, physical interactions occur between proteins andlipids and surfactants, forming molecular complexes (19). Investigation ofmixed adsorbed layers of protein and surfactant has revealed that polarlipids and surfactants can displace proteins from the interface. If theprotein is only partially displaced, the mixed adsorbed layer is often laterally


inhomgeneous, with a patchwise distribution of protein-rich and surfactant-rich regions. The displaced protein often remains attached to the interfaceas a wrinkled sheet in the protein-rich regions (‘‘orogenic’’ displacement,illustrated in Fig. 1) (20,21).

Another important aspect is protein hydrolysis, which leads to theformation of protein fragments, peptides, and free amino acids. Proteinhydrolysis usually occurs by proteolytic enzymes (e.g., during the prepara-tion of fermented products such as yogurt and cheese). It may also becarried out to improve the solubility of the protein (e.g., in the preparationof soy protein isolates or for the formulation of low-allergenic infant food).The emulsifying and foaming properties of peptides obtained by hydryolysisare strongly dependent on the molecular mass and composition. Usually,the surface activity and the film-stabilizing properties become much lowerwith decreasing molecular masses. As a rule of thumb, good surface-activeproperties were obtained for amphipolar peptides with molecular massesexceeding 2000Da and preferably containing disulfide-linked fragments(22,23).

C. Complexity of Processes Preceding Coalescencein Food Emulsions

In food emulsions, coalescence is usually not caused by a single process, but,instead, by a complex sequence of several processes preceding the final stepof coalescence. These preceding processes strongly vary among differentproducts and therefore cannot be generalized. However, some commontendencies can be distinguished.

Figure 1 Sketch of an oil–water interface with an adsorbed layer of protein, to

which a low-molecular-weight surfactant has been added, which displaces the

adsorbed protein layer by the ‘‘orogenic’’ mechanism, leading to patchwise

adsorption of protein and surfactant.


An important aspect in the application of proteins as emulsion stabi-lizers is that the molecular structure and properties of the molecules insolution may change during processing. For example, the solubility ishighly dependent on pH and, for most proteins, becomes very small inthe vicinity of their isoelectric pH. In many food products, the pH is loweredto around 4–5 during manufacture, which is close to the isoelectric pH ofmany proteins. Another example is that microbial contamination is oftenreduced in the last stage of production by a heat treatment (sterilization orpasteurization). This may lead to irreversible denaturation of the proteins,which, as a result, often become insoluble. The denatured protein may pre-cipitate as flocs and remain separate from the emulsion droplets or mayprecipitate onto the droplet surface, forming thick adsorbed layers or evenencapsulating the droplets in the precipitate (2).

The sensitivity to coalescence often changes during storage due toongoing chemical (e.g., oxidation), physical (e.g., creaming, crystallizationof lactose or inorganic salts), and enzymatic (e.g., protease and lipase activ-ity) transformations in the product. During handling, the emulsion isspread, stirred, mixed with other ingredients, acidified, heated, and exposedto air, all of which may lead to large structural and compositional changesin the emulsion, which may induce coalescence.

Finally, it is the complex behavior of the emulsion in the mouth thatdetermines its perception by the consumer. Several processes change thestructure of the emulsion and the release of flavor compounds (24), suchas mixing with saliva, melting of the gelled continuous phase, hydrolysis ofstarch by amylases in saliva (25), and possibly also release of fat from theemulsion and its adherence to the mucosa (4).

D. Aim of This Work

The complexity of the composition and structure of food emulsions andfoams complicates and understanding and adequate control the behaviorof these systems. The final step of coalescence can be caused by severalmechanisms and, as consequence, may reveal itself in many differentways, such as a coarsening of the droplet size distribution (dressings), for-mation of oil lenses on top of the emulsion (soups), and clumping of fat(cream). Because of the inherent complexity of food products and theirapplication, predictive tests to determine the stability of an emulsion tocoalescence are often not very reliable. These tests accelerate only one spe-cific mechanism, which is not necessarily the most important in the situationconsidered. For example, a centrifugation test (26) is not suitable to predictcoalescencemediated by the surface of expanding gas bubbles (Section IV. B).Clearly, there is a gap of knowledge in the realm of coalescence mechanisms


in food emulsions. The aim of the present work is give a general overview ofthe multitude of physical mechanisms in food emulsions that are collectivelyreferred to as ‘‘coalescence.’’

The organization of the main part of this chapter is as follows. Therole of the adsorbed layer in coalescence processes is discussed Section II.The structure and stability of thin films are related to the properties ofthe constituent adsorbed layers, making a comparison between surfactant-and protein-stabilized thin films. Section III describes homogeneous coales-cence. The main consecutive steps, droplet encounter and initiation of filmrupture, are distinguished, and for each step, various mechanisms will bediscussed. Finally, Section IV will focus on mechanisms of heterogeneouscoalescence, which involves the interaction with a third phase. Distinctionwill be made among coalescence intermediated by a solid and liquid sur-faces, coalescence induced by shearing in confined spaces, and coalescencecaused by puncturing of thin films by small particles.

II. ROLE OF THE ADSORBED LAYER IN COALESCENCE

A. Thin Films

Stability against coalescence is directly related to the properties of the pro-tective coating of adsorbed material at the droplet surface. This is becauseadsorbed layers on adjacent droplets repel each other (‘‘disjoining pres-sure’’), in this way keeping the internal phases of these droplets separated.This repulsion is usually of electrostatic steric origin. If the droplets attracteach other (e.g., by van der Waals forces) or are pushed together by anexternal force, a thin film is formed. The thickness of this thin film isdetermined by a balance of attractive and repulsive forces and mainlydepends on the range of the repulsive interactions (Fig. 2). For adsorbedlayers consisting of uncharged polar lipids, the repulsion is the short-rangeBorn repulsion. The thin film then consists of a bilayer of adsorbed materialand hardly contains free water in between the adsorbed layers. Such a thinfilm is called a Newton black film (NBF). For charged adsorbed layers oradsorbed layers with hydrophilic chains extending into the continuous liquid[an example of the latter is an adsorbed layer of b-casein in which chargedand hydrophilic N-terminal chain segment protrude into the water (27)], thedistance between the adsorbed layers remains larger and the thin films con-tain free water. Such a film is called a common black film (CBF).

Over the last decade, the understanding of thin films stabilized byproteins has substantially increased. At pH remote from the isoelectricpH, adsorbed protein layers are charged. If two droplets of a protein-stabilized emulsion are pushed together, the electrostatic repulsion between


these layers keeps the adsorbed layers separated, thus forming a CBF.Increasing the concentration of 1 : 1 electrolytes reduces the Debye length,thereby decreasing the range of double-layer repulsion and the thickness ofthe CBF. This also increases the stability of the NBF state with respect tothe CBF state and also reduces the energy barrier involved in the transitionof CBF to NBF. High-valency cations (Ca2þ, Al3þ) are even more effectivein this because they also lower the charge and surface potential by bindingto the adsorbed protein molecules (28–30).

As described in Section I. A, proteins change their conformation afteradsorption, leading to the formation of an elastic adsorbed film at the inter-face. At least for the whey proteins, this also leads to exposure of hydro-phobic amino acid residues and free sulfydryl groups to the water phase.When adsorbed layers on adjacent droplets are in contact, corresponding tothe NBF state, hydrophobic and disulfide bonds will then also be formedbetween these layers, strongly increasing the attractive interaction betweenthe adsorbed layers. The binding may become so strong that it becomeseffectively irreversible (15,30,31).

In dilute protein-stabilized emulsions at pH remote from the isoelectricpH, the electrostatic repulsion between the adsorbed layers keeps the drop-lets separated. Increasing the concentration of 1 : 1 electrolytes and espe-cially high-valency cations promotes droplet flocculation and stabilizes theNBF state with respect to the CBF state. Because the NBF state allowsthe formation of intermolecular bonds between adsorbed layers in contact,the binding between the adsorbed layers becomes enforced and dropletflocculation becomes irreversible (often referred to as ‘‘coagulation’’).

Figure 2 Transition between a common black film and a Newton black film,

induced by increasing the normal pressure onto the thin film or by reducing the

electrostatic repulsion between the films. Here, the thin films are stabilized by

negatively charged layers of adsorbed protein molecules.


In effect, on increasing the electrolyte concentration or approaching theisoelectric pH, the thin films between the flocculated droplets ‘‘jump’’from the CBF state into an irreversible NBF state (32).

B. Rupture of Thin Films

Film rupture occurs when a hole is formed in the thin film, with a diameterlarger than a critical size, so that the hole will grow spontaneously. The holecan be formed spontaneously or by mechanical forces pulling at the film.Film rupture related to spontaneous formation of a hole will be called‘‘spontaneous film rupture.’’ Film rupture induced by mechanical forceswill be called ‘‘induced film rupture.’’ Various mechanisms may initiatefilm rupture and the most important ones will be treated in Section III. A.The way by which adsorbed layers protect the thin film against coalescencedepends on the molecular properties of the adsorbed layer. Based onthe properties of adsorbed protein layers discussed in the previous sub-sections, we may expect large differences between the ways proteins andlow-molecular-weight surfactants to stabilize thin films.

Low-molecular-weight surfactants form mobile adsorption layers.Thin film stability is mainly due to a fast hole-repair mechanism (33), incombination with a preferred curvature of the adsorbed layer that opposesthe mean curvature of a neck of liquid connecting the liquid volumes onboth sides of the thin film (34). If the surfactant is charged or exposes ahydrophilic chain into the solution, a relatively large distance is maintainedbetween the two sides of the thin film due to electrostatic or steric repulsion,respectively, which improves the stability of the thin film.

Proteins, on the other hand, form thick, cohesive, and elastic adsorbedlayers and the molecules in the adsorbed layer are relatively immobile (35).These properties afford a strong barrier against film rupture under quiescentconditions (36). However, on large deformation, the adsorbed layers mayfracture (Section I.B) and this may have important consequences for thestability of the thin films (see Section III. A. 1). As discussed in the previoussection, on close contact between adsorbed layers, strong hydrophobic andcovalent bonds can be formed between the adsorbed layers, leading tocoagulation of the droplets. The presence of these bonds makes the emulsionsensitive to a specific mechanism of coalescence that will be discussion inSection III.A.3.

III. HOMOGENEOUS COALESCENCE

Commonly, coalescence involves the process of merging of two emulsiondroplets, dispersed in a continuous liquid. This process of coalescence


generally comprises two main consecutive steps: (1) encounter of dropletsurfaces and the formation of a thin film and (2) initiation of thin-filmrupture. The first step corresponds to the flocculation (reversible) andcoagulation (irreversible) of the emulsion droplets (steps A and B inFig. 3). The second step corresponds to the actual irreversible process ofcoalescence (steps C and D in Fig. 3). A detailed treatment on this subject interms of the characteristic times for flocculation, coagulation, and filmrupture has been given by Dukhin et al. (37). Which of the two mainsteps will be rate determining not only depends on the properties of theadsorbed layer but also on many other aspects, such as the droplet-sizedistribution, the volume fraction of droplets and the presence of liquidflow. These two main steps will be discussed in Sections III.A and III.B.

A. Initiation of Thin-Film Rupture

Film rupture is the final step of the coalescence process. It is initiated by theformation of a hole larger than a critical size in the thin film, leading to themerging of two droplets into one (Fig. 3, step C). Hole formation may havea number of causes, of which vacancies in the adsorbed layer, spontaneouspassage formation and rupture by film stretching are the most important.

1. Rupture by the Presence of Vacancies

This mechanism prevails if the adsorption is below saturation adsorption,where part of the droplet surface remains uncovered by an adsorbed layer. Ifthe adsorbed molecules attract each other, a two-dimensional phase separa-tion between closed-packed molecules and empty regions (‘‘vacancies’’) willtake place (Fig. 4). Film rupture is initiated if sufficiently large vacancies

Figure 3 Basic coalescence process involving two droplets: (A) encounter of

droplets; (B) thin-film formation; (C) thin-film rupture; (D) merging of droplets.


adjoin on both sides of the thin film. The lifetime of such a film thendepends on nucleation and growth of these vacancies. This mechanismwas modelled by Kashchiev and Exerova (38) and Exerova and Nikolova(39) to calculate the lifetime of surfactant bilayer films (which are equivalentto Newton black films).

Slighly below saturation adsorption of the surfactants in the adjacentmonolayers, the model predicts a sharp transition toward very high stabilityof the bilayer. Also, for protein-stabilized films, a similar sharp transition inthin-film stability was found when the adsorption density reached saturationadsorption (40,41).

2. Rupture by Spontaneous Passage Formation

Here, a ‘‘passage’’ or ‘‘neck’’ is defined as a canal of the dispersed liquidpassing through the sheet of thin film, in which the surface between thecanal and the thin-film liquid is covered by an adsorbed layer. In thisway, the surface of the passage forms a continuous bridge between thesurfaces of the adjoining emulsion droplets (Fig. 5). Theories for the initia-tion of rupture by spontaneous passage formation have been developedmainly for surfactant-stabilized emulsions.

De Vries (42) described initiation of rupture of common black films bycalculating the energy barrier related to the area increase needed to form apassage through the layer of liquid water. Related to this is a theoreticaltreatment by Harbich et al. (43) for the energy barrier involving passageformation in vesicles and lamellar phases. Later, Kozlov and Markin (44)extended this treatment by incorporating curvature components of the sur-face tension to describe the energy effect involved in the surface-topologicaltransition of merging two droplets into one. Recently, Kabalnov and

Figure 4 Sketch of a thin film between emulsion droplets, with vacancies in the

adsorbed layer due to unsaturation of the adsorbed layers.


Wennerstrom (34) further developed this approach for the formation of apassage in thin films between emulsion droplets stabilized by low-molecular-weight surfactants for temperatures close to the phase-inversion tempera-ture. In this ‘‘oriented-wedge theory,’’ the formation of a passage through athin film of the continuous liquid is strongly inhibited if the surfactant tendsto form micelles in the continuous phase (Fig. 6). This mechanism seems tobe unimportant for emulsions stabilized by only protein, because of thethickness and hydrophilicity of the adsorbed layer (40).

3. Rupture by Film Stretching

Thin films may rupture when they are stretched at such a speed that thereduction of the adsorption density of the stabilizing surface-active materialcannot be restored by adsorption from the bulk or by lateral mobility ofadsorbed material. Stretching then leads to depletion of adsorbed materialin the adsorbed layers, which may induce rupture by the presence ofvacancies (Section III.A.1). This may occur for emulsion stabilizersthat form relatively immobile films and are dissolved in the continuousliquid, which makes transport toward the middle of the thin film aslow process. These requirements are met for adsorbed protein layers(Section I.B). For thin films stabilized by cohesive layers of adsorbed

Figure 5 Sketch of rupture of a thin film between adjacent emulsion droplets, due

to the formation of a passage through the thin film.


protein molecules (Section I.B), we may also expect that initiation of rupturecorresponds to fracturing of these adsorbed layers.

An additional requirement of this film rupturing mechanism is thatexternal forces exerted on the emulsion lead to stretching of the thin films.This will only occur if external forces exerted on an emulsion are transferredonto the thin film in a direction parallel to the thin film. As a consequence,this mechanism will not be effective in relatively dilute, unaggregatedemulsions under quiescent or gentle-flow conditions.

Rupture by film stretching was shown to occur in highly concentratedemulsions stabilized by protein (40,45–47). Emulsions are said to be highlyconcentrated whenever the volume fraction of the droplets exceeds the limitof close packing (Fig. 7). The starting point for explaining the experimentalresults is the hypothesis that film stretching occurs in these systems if exter-nally applied stresses cannot relax sufficient by flow in the emulsion (46).

Figure 6 Oriented-wedge model for film-stabilizing properties of low-molecular–

weight surfactants. The surfactant molecules are depicted as wedges, with the

broader side oriented either to the oil phase (upper diagram) or the aqueous phase

(lower diagram). The upper case corresponds to low-hydrophilic–lipophilic-balance

(low-HLB) surfactants, which are able to form micelles in the oil phase and easily

form a passage through an aqueous thin film, therefore not suited to stabilize oil-in-

water emulsions. The lower diagram corresponds to high-HLB surfactants, which

are able to form micelles in the aqueous phase and inhibit the formation of a passage

through the aqueous film, therefore suited to stabilize oil-in-water emulsions.


This occurs when slip between the droplet surfaces is inhibited by the pres-ence of shear-resisting connections between the adsorbed layers of adjacentdroplets (47). Van Aken et al. (46) suggested that the formation of shear-resisting connections is closely related to the transition of CBF to NBF thatwas discussed in the context of droplet aggregation in dilute emulsionsstabilized by whey protein in Section II. B. Nuclear magnetic resonance(NMR) mobility measurements of water molecules pointed out that astate corresponding to a Newton black film (essentially no free waterbetween the adsorbed layers forming the thin film) can indeed be obtainedin a highly concentrated emulsion stabilized by b-lactoglobulin (48).

Similarly to the bonds formed in dilute emulsions stabilized by wheyprotein, the shear-resisting connections are probably formed by stronglyattractive short-range forces such as hydrophobic interaction (47). Theywere shown to form between adsorbed protein layers of both b-lactoglobu-lin and b-casein when the droplet surfaces were pushed together with suffi-ciently large pressure (46). It was argued that this critical pressure is relatedto a pressure-induced jump over the kinetic barrier that is related to thetransition of CBF to NBF. This was supported by kinetic measurementsthat showed that the rate of formation of shear-resisting connectionsincreases as a function of the applied pressure and the ionic strength, andwas strongly enhanced in the presence of calcium ions (47).

This mechanism was shown to be relevant also is a more dilute emul-sion of irreversibly aggregated droplets, where deformation of dropletaggregates by shear stresses may cause coalescence of droplets within theaggregate (Fig. 8) (47).

Figure 7 Formation of a highly concentrated emulsion by removal of the aqueous

phase.


B. Encounter of Droplet Surfaces and the Formation of a Thin Film

Before a thin film is formed, two consecutive steps have to take place(Fig. 3). First, the emulsion droplets have to come close together on acolloidal scale-droplet encounter, step A in Fig. 3) and second, a thin filmmust be formed on a molecular scale (outflow of the continuous phase fromthe spacing between the droplet and thin-film formation, step B in Fig. 3).

1. Droplet Encounter Under Quiescent and Gentle-Flow Conditions

Droplet encounter may occur through diffusion (in analogy to fast periki-netic flocculation (49,50)) but is strongly accelerated by convection becausethis increases the encounter rate of the emulsion droplets (in analogy to fastorthokinetic flocculation (51)) and very weak repulsion between the dropletsurfaces may be overcome by the shearing force.

The rate of droplet encounters is greatly increased by concentratingthe emulsion either by an artificial procedure such as reverse osmosis orcentrifugation or spontaneously by creaming (52). The ultimate situation isthe previously discussed case of highly concentrated emulsions, in which thedroplets are in continuous contact with each other (Fig. 7).

When the surfaces are sufficiently close together, they start to interact,first by hydrodynamic interaction, which retards the outflow of liquid(53–56), but at sufficiently short separation also by dispersion forcesbetween the liquid interiors of the droplets and short-range forces betweenthe adsorbed layers (57). The outflow of liquid becomes retarded if thehydrodynamic repulsion between the approaching droplets deforms thedroplets, leading to the formation of a (dimpled) thin film (57). The retarda-tion is enhanced by the elasticity of the adsorbed layers, because this resiststhe outflow of liquid by the coupling of liquid flow and surface flow (58).

Figure 8 Deformation of a droplet aggregate by external forces, inducing

coalescence by stretching of thin films between the droplets because slip between

the adsorbed layers of the thin films is inhibited.


This retardation is of importance mainly if the droplets are relatively largethe surface tension is relatively low and under highly dynamic condition,such as during emulsification (58).

Finally, thin spots can form in the remaining liquid film by a spinodalthinning mechanism, described by Scheludko (59) and Vrij (60). Thismechanism describes local thinning of a CBF caused by thermal undulationsof the film thickness and is commonly accepted as one of the main steps inthe process of spontaneous film rupture (61). This mechanism is retarded byincreasing the viscosity of the continuous liquid and surface dilational elas-ticity of the adsorbed layers. Already at a very low surface elastic modulus( 1mN/m), the surface dilational elasticity strongly suppresses the spino-dal thinning mechanism (62). This is the case for adsorbed protein layers,because the surface dilational elasticity is usually in the range 20–80mN/m.Therefore, thinning of protein-stabilized films is strongly retarded and thespinodal thinning process is strongly damped, predicting that also the tran-sition CBF to NBF is strongly retarded for protein-stabilized emulsions, aswas indeed observed experimentally (35,63).

2. Droplet Encounter in Turbulent Flow

In turbulent flow, droplets bounce together by inertial forces, which maylead to large droplet deformation and to forces stretching the adsorbedlayers and thin films between colliding droplets. Film rupture may thenoccur by similar mechanisms as described in Section III. A.3. As a firstapproach to describing homogeneous coalescence induced by turbulentflow, we will estimate the conditions needed for the occurrence of coales-cence in highly turbulent flow. In turbulent flow, the outflow of the con-tinuous phase from the spacing between the droplets may be greatlyenhanced by the inertial between the droplets when they are bouncingagainst each other. Walstra and Smulders (64) estimated the magnitude ofthis pressure from Kolmogorov theory as

�p ¼ "2=3d2=3�1=3 ð1Þ

where " is the power density of the turbulent flow, d is the length-scaledistance considered, for which we take the droplet diameter, and � is thedensity of the emulsion. The outflowing liquid exerts a shear force onto theadsorbed layers of the thin film, tending to stretch them. This may lead tofilm rupture by film stretching as described in Section III.A.3 (Fig. 9). Thiscan be quantified by taking into account the surface tension gradient �galong the surfaces of the thin film, which for immobile adsorbed layers


will be magntitude (64)

�� 14ð�pÞd ð2Þ

Combination with Eq. (1) yields

�� ¼ 14"2=3d5=3�2=3 ð3Þ

For �¼ 1000 kg/m3 and d¼ 10�6m, we find �g¼ 2.5� 10�8"2/3. Foradsorbed proteins, the maximum reduction of the equilibrium surface ten-sion is limited to approximately 0.03N/m, which is related to the occurrenceof collapse of the adsorbed layer above this value. This value may thereforebe taken as an estimate of �g, although for the very short time scale ofdroplet impact and the resulting film stretching, higher values may be attain-able. Moreover, the actual rupture mechanism will probably be related tofracture of the adsorbed layer rather than collapse. Nevertheless, taking this

Figure 9 Coalescence induced by film stretching of droplets impacting in highly

turbulent flow. The second drawing shows the outflow of continuous liquid from the

spacing between the approaching droplets, as shown by flow lines, which drags along

the adsorbed layers at the droplet interfaces. This produces a spot with high surface

tension and low adsorption density in the thin film, as indicated by the asterisk,

initiating film rupture and coalescence.


estimate for �g, coalescence is expected if ">1.3� 109W/m3. This is muchhigher than highest value for the power density that can be obtained by atoothed colloid mill in highly viscous emulsions (for this situation, a max-imum power density of approximately 107W/m3 was reported by Karbsteinand Schubert (65)). According to Walstra and Smulders, an Ultra-Turraxtype of machine and high-pressure homogenisers can produce power den-sities up to 1010 and 1012W/m3, respectively (64), which should be sufficientfor this coalescence mechanism.

Little experimental evidence on this coalescence mechanism can befound in the literature. One of the few studies on this subject was byChen et al. (66), who showed an increased rate of coalescence underhighly turbulent flow conditions. However, it cannot be excluded thatcoalescence was caused by the introduction of air (which may lead to air-intermediated coalescence, discussed in Section IV.B) or occurred atthe surface of the impeller blade (which may lead to surface-intermediatedcoalescence, discussed in Section IV. A).

For surfactant-stabilized emulsions, surface-tension gradients maybe suppressed by fast exchange of the surfactant from solution.However, in the thin-film region, supply of surfactant from the dispersedphase is much more efficient than from the continuous phase. Therefore,if the surfactant is preferentially dissolved in the dispersed phase, itis more difficult to induce surface-tension gradients and vacancies in thethin-film region, and film thinning will proceed faster than if the surfac-tant is preferentially dissolved in the continuous phase. However, ifthe preferential partitioning of the surfactant in the dispersed phase iscaused by micelle formation in this phase, the oriented-wedge theorywould predict fast film rupture by passage formation in this case(Section III. A. 2).

IV. SURFACE-INTERMEDIATED COALESCENCE

In some cases, a third phase intermediates in the coalescence of emulsiondroplets. This class of coalescence mechanisms has received relatively littleattention in the literature. Based on general considerations, the critical stepmust be the rupturing of a thin water film between the droplet and the thirdphase. The third phase may be solid (e.g., the wall of the processing equip-ment or the wall of the container in which the emulsion is stored, or a gas,such as the headspace in the container or air bubbles beaten in duringpumping or whipping). A special, very important case is the situationwhere the third phase is formed by small solid particles. These cases willbe treated in the following subsections.


A. Coalescence Intermediated by a Solid Surface

Although not much is known about this mechanism, we may expect thatcoalescence intermediated by a solid surface is related to spreading of thedispersed phase onto this surface (Fig. 10). A crucial requirement for thismechanism must be that the dispersed phase (partially) wets the surface; thecontact angle should be lower than 90�or so. The kinetics of this mechanismwill be determined by the rupture of the thin (aqueous) film between thedroplet and the solid surface, which can be understood as the dewetting ofthe thin film of the continuous liquid at the solid surface. Dewetting hasbeen described by spinodal dewetting (67) and by nucleation by nanoscalegas nuclei, as first suggested by Derjaguin and Gutop (68).

It has been suggested that coalescence intermediated by a solid surfacemay occur at the surface of the impeller blades in an agitated vessel, in thecontext of phase inversion of liquid dispersions (69). The process would beimportant if the dispersed liquid has a higher density than the continuousliquid, so that the drops are able to reach the impeller surface by inertialimpaction, where they may spread at the impeller surface or not, dependingon the wetting conditions. Also, for emulsification in a colloid mill, thespreading behavior of the dispersed phase onto the wall of the colloidmill was shown to be important for the efficiency of the emulsificationprocess (70).

B. Coalescence Intermediated by a Liquid Interface

Coalescence intermediated by a liquid interface (the interface between agas and a liquid or between two immiscible liquids) resembles that at asolid surface, but is different because interfacial flow and area dilations arepossible (Fig. 11). A requirement is again that the dispersed phase mustspread upon the liquid interface, which is the case for pure triglycerides atpure air–water interfaces. Under quiescent conditions, the liquid interfacewill be covered by an adsorbed layer of surface-active material supplied bythe solution. Emulsion droplets that approach the interface will be repelledby this layer, and on close encounter, a thin film will be formed. The rate

Figure 10 Coalescence intermediated by adsorption and (partial) wetting of

droplets on the surface of a solid substrate.


of insertion and spreading of these droplets at the interface will thendepend on the lifetime of this thin film. If the interface is expanded, theadsorption density at the interface will decrease by an amount thatdepends on the rate of adsorption of the surface-active material (Fig. 12).Higher expansion rates result in lower adsorption densities and highersurface tensions of the air–water interface. Moreover, if a cohesiveadsorbed layer is formed, expansion of this layer may even result in theformation of bare cracks at the air–water interface by fracturing of thislayer. In this way, expansion of the air–water interface will lower stabilityof the thin film between the expanding surface and the emulsion droplet,promoting droplet insertion.

Coalescence intermediated by air was studied at an expanding inter-face by Hotrum et al. (71) for emulsions stabilized by whey proteinand b-lactoglobulin. It was shown that droplet spreading occurred if theair–water interface was expanded at such a rate that the surface tension ofthe interface exceeded a critical value. This critical value was shown to

Figure 11 Coalescence intermediated by adsorption and spreading of the droplets

at an expanding interface. (A) Droplets do not enter the air–water interface because

of the kinetic stabilization by an air–water–oil thin film; (B) expansion of the air–

water interface lowers the adsorption density at the air–water interface, leads to

convective transport of droplets toward the air–water interface and increases the

surface tension of the air–water interface to a value that enables spreading of the

liquid oil onto this interface; (C) after spreading, the oil film contracts and breaks up

into oil lenses.


correspond to the criterion of spreading of the oil at the air–water interface.Insertion of droplets without spreading should, in principle, be possible atlower expansion rates; however, this was not observed, probably because itis strongly suppressed by the high stability of the thin films between theemulsion droplets and the interface.

A self-amplifying mechanism for droplet spreading at the expandinginterface was suggested by Hotrum et al (71). Spreading of oil at anair–water interface is a very fast process, with high expansion rates at theoil–water interface of the spreading emulsion droplet. The rapidly expand-ing oil–water interface may attract emulsion droplets from the underlyingsolution by convection, and if the spreading oil film is not sufficientlyquickly covered by a protecting adsorbed layer, these droplets immediatelycoalesce with the spreading oil film. In this way, the spreading oil film issupplied with additional oil from emulsion droplets coalescing with thespreading film, amplifying the process.

After expansion, the spread oil film may be reintroduced into theemulsion in the form of droplets if the air–water interface disappears. Infoamed emulsions, this may, for example, occur by coalescence or dispro-portionation of the gas bubbles. A well-known practical application of thisprocess is the conventional process of butter manufacture by churning,where coalescence of milk fat globules and subsequent clumping tobutter granules is intermediated by expanding and disappearing air–waterinterfaces (1).

Figure 12 Coalescence intermediated by adsorption and spreading of the droplets

at an expanding interface. Sketch showing how expansion of the air–water interface

may reduce the stability of the air–water–oil thin film, promoting insertion of

droplets at the air–water interface, which may initiate subsequent spreading of

droplets. (A) quiescent air–water interface; (B) expanding air–water interface, with

reduced adsorption density of surface-active material; (C) situation just after

insertion: a lens is formed and surface-active material previously adsorbed on the

droplet is transferred to the air–water interface.


Coalescence intermediated by a liquid interface may also be of greatimportance in several other situations. For example, during whipping ofcream, dilating bubble surfaces may cause in insertion and (partial) spread-ing of milk fat globules. When an emulsion is poured, the air–water interfacewith area A will expand. The expansion rate (d lnA/dt) depends on thepouring rate and presence of surface-active material and will be on theorder of 0.1 s�1, similar to expansion rates measured in overflowing cylinders(72) or canals (73). During pumping, air bubbles may be beaten into theemulsion and be strongly deformed, leading to local high expansion ratesof the bubble surfaces. On depressurizing or heating an emulsion, dissolvedgas may be released as growing gas bubbles. The same may happen if gas isproduced by (bio-)chemical reactions (e.g., the release of carbon dioxideduring fermentation). In all cases, this may cause insertion and spreadingof the emulsion droplets, by an extent that depends on the expansion rate ofthe interface and the rate of supply of film-stabilizing material from solution.

C. Coalescence by Shearing in Confined Spaces

Probably of great practical importance is the behavior of emulsions whenconfined to a thin layer between solid surfaces in relative motion to eachother (Fig. 13). It is thought that such a situation occurs during oral process-ing, where the emulsion lubricates the movement between the tongue andpalate (74). However, also during processing (e.g., homogenization in acolloid mill) and handling (e.g., spreading of a food emulsion with aknife), similar processes may occur. Obviously, conditions of high shearcan easily be obtained in these configurations, but a special situationoccurs when the distance between the shearing plates is the same order ofmagnitude as the droplet size or droplet size aggregates or is smaller thanthis. This is a largely unexplored field of thin-film tribological behavior,governed by large deformations and breakup of droplets and droplet

Figure 13 Coalescence induced by shearing in confined spaces. To illustrate

possible processes, a combination of homogeneous and surface-intermediated

coalescence is sketched for a droplet aggregate sheared between two closely spaced

surfaces.


aggregates, droplet coalescence, and the interaction of the droplets with thesolid surfaces (wetting, contact angles).

D. Rupture by Penetration of Particles

Film rupture by penetration of particles can be viewed as a special case ofsurface-intermediated coalescence (Fig. 14). It is a very commonly encoun-tered mechanism. Particles that are fully wetted by either the dispersedliquid or the continuous liquid may puncture the thin film due to mechanicalforces. However, often the particles are partially wetted by the other phase(finite contact angle at the oil–water interface), in which case the crystalshave the tendency to reside at the oil–water interface and bridge the thinfilm. The particles may originate from the aqueous phase (e.g., calciumphosphate, lactose) or the oil phase (fat crystals).

Rupture by particles dispersed in the continuous liquid has beenstudied for foams in relation to antifoam agents (75), but much less foremulsions. De Gennes (76) has described a mechanism for coalescencecaused by film rupture by penetration of particles. In this model, a singleparticle that remains emerged in the continuous phase may lead to asequence of film ruptures that finally results in a few large dropletsembedded in remaining uncoalesced droplets. An example of coalescencecaused by penetration of thin films by water-soluble sodium nitrate crystalsis shown in Fig. 15. Here, a highly concentrated emulsion, stabilized bywhey protein isolate, was dried by exposure to the air. The emulsion wasquickly destabilized and separated oil as soon as sodium nitrate, dissolved inthe aqueous phase, started to crystallize.

Figure 14 Coalescence intermediated by film penetration by particles suspended in

(a) the dispersed oil and (b) the continuous water. The particles may or may

not reside at the oil–water interface because of partial wetting by both liquids

(as explained in the text).


Figure 15 A highly concentrated oil-in-water emulsion (droplet diameter¼ 1 mm,

volume fraction of oil¼ 0.9), stabilized by whey protein isolate and containing 0.1M

sodium nitrate dissolved in the aqueous phase. Visible are sodium nitrate crystals

that were formed when the emulsion was dried by exposure to the air. The formation

of these crystals lead to rapid separation of oil from the highly concentrated

emulsion, which did not occur if a similar highly concentrated emulsion without

sodium nitrate was dried. Individual emulsion droplets or separated oil cannot be

observed because of the low optical contrast. The drawing shows a sketch of the

situation.


An important example of rupture by particles embedded in the dis-persed phase is the phenomenon of ‘‘partial coalescence,’’ which has beenstudied extensively for dairy-related emulsions (77,78). Here, the particlesare fat crystals that are present in the oil droplets at the temperature ofapplication. Dependent on the presence of surfactants or polar lipids and onthe mobility of the fat crystals inside the droplet, these fat crystals may moveto and reside at the oil–water interface, finally leading to rupture of the thinfilms. Partial coalescence of emulsions of partially crystallized fat leads tothe formation of clumps of fat globules, which retain part of their originaldroplet structure because the fat crystals inhibit the flow of two sphericaldroplets into one (step D in Fig. 3). In this way, networks of clumped fatglobules can be formed, which are important for structuring products suchas whipped cream and ice cream.

V. CONCLUSIONS AND PERSPECTIVE

A large number of coalescence processes may be responsible for the coars-ening of the droplets in a food emulsions. Fundamental knowledge of theseprocesses is needed in order to distinguish them and to take adequate actionto control them. The intention of this work was to give an overview ofpossible mechanisms, some of which are already known in the literature,but most of them are hardly studied. The present overview may initiate newresearch in this complex field of coalescence in food emulsions.

REFERENCES

1. H. Mulder and P. Walstra, The Milk Fat Globule, Pudoc, Wageningen, The

Netherlands, 1974.

2. P. Walstra and R. Jenness, Dairy Chemistry and Physics, Wiley, New York,

1984.

3. G. A. van Aken, Aeration of emulsions by whipping, Colloids Surfaces A 190,

333–354 (2001).

4. A. Akashe, D. Das, and M. Miller (inventors), Mesophase-stabilized emulsions

and dispersions for use in low-fat and fat-free food products, U.S. patent No.

6068876, 2001.

5. C. R. T. Brown, J. E. Homan, M. G. Jones, M. E. Malone, and I. T. Norton,

(inventors), Non-frozen, low-fat food emulsions and processes thereof, U.S.

patent No. 2001041207, 2001.

6. P. E. A. Smulders, Formation and stability of emulsions made with proteins and

peptides, Thesis, Wageningen University, Wageningen, The Netherlands, 2000.

7. G. Schuster, Emulgatoren fur Lebensmittel, Springer-Verlag, Berlin, 1985.


8. N. J. Krog, Food emulsifiers and their chemical and physical properties,

in Food Emulsions, K. Larsson and S.E. Friberg,(eds.), K. Marcel Dekker,

New York, 1990 pp.127–180.

9. K. D. Schwenke, Proteins: some principles of classification and structure, in

D. Proteins at Liquid Interfaces, D. Mobius, and R. Miller, (eds.), Elsevier,

Amsterdam, 1998, pp.1–50.

10. F. MacRitchie, Reversibility of protein adsorption, in Proteins at Liquid Inter-

faces, D.Mobius, and R.Miller, (eds.), Elsevier, Amsterdam, 1998, pp. 149–177.

11. M. Corredig and D. G. Dalgleish, A differential microcalorimetric study of

whey proteins and their behaviour in oil-in-water emulsions, Colloids Surfaces

B 4, 411–422 (1995).

12. Y. Fang and D. G. Dalgleish, Conformation of b-lactoglobulin studied by

FTIR: Effect of pH, temperature, and adsorption to the oil-water interface,

J. Colloid Inter. Sci.196, 292–298 (1997).

13. E. Dufour, M. Dalgalarrondo, and L. Adam, Conformation of b-lactoglobulinat an oil/water interface as determined from proteolysis and spectroscopic

methods, J. Colloid Inter. Sci. 207, 264–272 (1998).

14. E. Dickinson and Y. Matsumura, Time-dependent polymerisation of

b-lactoglobulin through disulphide bonds at the oil–water interface in

emulsions, Int. J. Biol. Macromol. 13, 26–32 (1991).

15. D. J. McClements, F. J. Monahan, and J. E. Kinsella, Disulfide bond

formation affects stability of whey-protein isolate emulsions, J. Food Sci. 58,

1036–1039 (1993).

16. E. Dickinson and Y. Matsumura, Protein at liquid interfaces: role of the molten

globule state, Colloids Surfaces B 3, 1–17(1993).

17. A. Martin, M. Bos, M. A. Cohen Stuart, and T. van Vliet, Stress–strain curves

of adsorbed protein layers at the air/water interface measured with surface

shear rheology, Langmuir 18, 1238–1243 (2002).

18. P. J. Wilde and D. K. Sarker, Enhancement of bubble surface elasticity by

crosslinking agents and their effects on protein foam stability, in Bubbles in

Food. (G. M. Campbell, C. Webb, S. S. Pandiella, and K. Niranjan eds.),

American Association of Cereal Chemists, St. Paul, MN, 1999, pp.145–152.

19. B. Ericsson, Lipid–protein interactions, in Food Emulsions K. Larsson and

S. E. Friberg, eds., Marcel Dekker, New York, 1990, pp. 181–201.

20. A. R. Mackie, P. A. Gunning, P. J. Wilde, and R. J. Morris, Competitive

displacement of b-lactoglobulin from the air/water interface by sodium dodecyl

sulfate, Langmuir 16, 8176–8181 (2000).

21. A. R. Mackie, P. A. Gunning, P. J. Wilde, and V. J. Morris, Orogenic dis-

placement of protein from the oil/water interface, Langmuir 16, 2242–2247

(2000).

22. P. W. J. R. Caessens, S. Visser, H. Gruppen, G. A van Aken, and A. G. J.

Voragen, Emulsion and foam properties of plasmin derived b-casein peptides,

Int. Dairy J. 9, 347–351 (1999).

23. P. W. J. R. Caessens, W. F. Daamen, H. Gruppen, S. Visser, and A. G. J.

Voragen, b-Lactoglobulin hydrolysis. 2. Peptide identification, SH/SS


exchange, and functional properties of hydrolysate fractions formed by the

action of plasmin, J. Agric. Food Chem. 47, 2980–2990 (1999).

24. S. Odake, J. P. Roozen, and J. J. Burger, Effect of saliva dilution on the release

of diacetyl and 2-heptanone from cream style dressings, Nahrung 42, 385–391

(1998).

25. L. M. Beidler, Saliva. Its functions and disorders, in Handbook of Olfaction and

Gustation (R.L. Doty, ed.), Marcel Dekker, New York, 1995, pp. 503–519.

26. R. D. Vold and R. C. Groot, An ultracentrifugal method for the quantitative

determination of emulsion stability, J. Phys. Chem. (Ithaca) 66, 1969–1975

(1962).

27. P. J. Atkinson, E. Dickinson, D. S. Horne, and R. M. Richardson, Neutron

reflectivity of adsorbed beta-casein and beta-lactoglobulin at the air/water

interface, J. Chem. Soc. Faraday Trans. 91, 2847–2854 (2001).

28. A. Kulmyrzaev, R. Chanamai, and D. J. McClements, Influence of pH and

CaCl2 on the stability of dilute whey protein stabilized emulsions, Food Res.

Int. 33, 15–20 (2000).

29. S. P. F. M. Roefs and H. A. Peppelman, Aggregation and gelation of whey

proteins: Specific effect of divalent cations? in Food Colloids: Fundamentals of

Formulation, (E. Dickinson and R. Miller, eds.), Royal Society of Chemistry,

Cambridge, 2001, pp. 358–368.

30. O. D. Velev, B. E. Campbell, and R. P. Borwankar, Effect of calcium ions and

environmental conditions on the properties of b-casein stabilized films and

emulsions, Langmuir 14, 4122–4130 (1998).

31. S. Damodaran and K. Anand, Sulfhydryl–disulfide interchange-induced

interparticle protein polymerisation in whey protein-stabilized emulsions and

its relation to emulsion stability, J. Agric. Food Chem. 45, 3813–3820 (1997).

32. D. J. McClements, Food Emulsions: Principles, Practice and Techniques, CRC

Press, Boca Raton, FL, 1999.

33. D. C. Clark, R. Dann, A. R. Mackie, J. Mingins, A. C. Pinder, P. W. Purdy,

E. J. Russell, L. J. Smith, and D. R. Wilson, Surface diffusion in sodium dodecyl

sulfate-stabilized thin liquid films, J. Colloid Inter.Sci. 138, 195–206 (1990).

34. A. Kabalnov and H. Wennerstrom, Macroemulsion stability: The oriented

wedge theory revisited, Langmuir 12, 276–292 (1996).

35. D. C. Clark, M. Coke, A. R. Mackie, A. C. Pinder, and B. Wilson, Molecular

diffusion and thickness measurements of protein-stabilized thin liquid films,


36. B. Biswas and D. A. Haydon, The coalescence of droplets stabilised by

viscoelastic adsorbed films, Kolloid Z. 185, 31–38 (1962).

37. S. S. Dukhin, J. Sjoblom, D. T. Wasan, and Ø. Sæther, Coalescence coupled

with either coagulation or flocculation in dilute emulsions, Colloids Surfaces A

180, 223–234 (2001).

38. D. Kaschiev and D. Exerova, Nucleation mechanism of rupture of Newtonian

black films, J. Colloid Interf. Sci. 77, 501–511 (1980).

39. D. Exerova and A. Nikolova, Short-range molecular interactions and stability

of amphiphile bilayers, J. Dispers. Sci. Technol. 18, 683–697 (1997).


40. G. A. van Aken and F. D. Zoet, Coalescence in highly concentrated coarse


41. S. Tcholakova, N. D. Denkov, I. B. Ivanov, and B. Campbell, Coalescence in

b-lactoglobulin stabilized emulsions: Effects of protein adsorption and drop

size, Langmuir 18, 8960–8971 (2002).

42. A. J. de Vries, Foam stability, Rec. Trav. Chim. 77, 383–399 (1958).

43. W. Harbich, R. M. Servuss, and W. Helfrich, Passages in lecithin-water

systems, Z. Naturforsch. 33a, 1013–1017 (1978).

44. M. M. Kozlov and V. S. Markin, Biofizika 28, 242 (1983).

45. G. A. van Aken and T. van Vliet, Mechanisms of coalescence in highly

concentrated protein-stabilized emulsions, in Food Colloids Fundamentals of

Formulation (E. Dickinson and R. Miller, eds.), The Royal Society of


46. G. A. van Aken, Flow-induced coalescence in protein-stabilized highly

concentrated emulsions, Langmuir 18, 2549–2556 (2002).

47. G. A. van Aken and T. van Vliet, Flow-induced coalescence in protein-

stabilized highly concentrated emulsions: role of shear resisting connections

between the droplets, Langmuir 18, 7364–7370 (2002).

48. G. A. van Aken, F. D. Zoet, and J. Diederen, Composition of thin films

between emulsion droplets stabilized by protein, as measured in highly

concentrated emulsions, Colloids Surfaces B 26, 269–279 (2002).

49. M. von Smoluchowski, Phys. Z 17, 557–585 (1916).

50. M. von Smoluchowski, Z. Phys. Chem. 92, 129 (1917).

51. T. G. M. van de Ven, Colloidal Hydrodynamics, Academic Press, London, 1988.

52. S. Kumar, G. Narshimhan, and D. Ramkrishna, Coalescence in creaming

emulsions. Existence of a pure coalescence zone, Ind. Eng. Chem. Res. 35,

3155–3162 (1996).

53. B. V. Derjaguin and V. M. Muller, Dokl. Akad. Nauk. SSSR (Eng.) 176, 738

(1967).

54. E. P. Honig, G. J. Roebertsen, and P.H. Wiersema, Effect of hydrodynamic

interaction on the coagulation rate of hydrophobic colloids, J. Colloid Interf.

Sci. 36, 97–109 (1971).

55. L. A. Spielman, Viscous interactions in Brownian coagulation, J.Colloid Interf.

Sci. 33, 562–671 (1970).

56. S. Hartland, Coalescence in dense-packed dispersions, in Thin Liquid Films

(I. B. Ivanov, eds.), Marcel Dekker, New York, 1988, pp 663–766.

57. O. D. Velev, K. D. Danov, and I. B. Ivanov, Stability of emulsions under static

and dynamic conditions, J. Dispers. Sci. Technol. 18, 625–645 (1997).

58. P. Walstra and P. E. A. Smulders, Making emulsions and foams :an overview,

in Food Colloids; Proteins, Lipids and Polysaccharides (E. Dickinson and

B. Bergenstahl, eds.), The Royal Society of Chemistry, Cambridge 1997,

pp. 367–381.

59. A. Scheludko, Thin liquid films, Adv. Colloid Interf. Sci. 1, 391–464 (1967).

60. A. Vrij, Possible mechanism for the spontaneous rupture of thin, free liquid

films, Discuss. Faraday Soc. 42, 23–33 (1966).


61. I. B. Ivanov and D. S. Dimitrov, Thin film drainage, in Thin Liquid Films:

Fundamentals of Application (I. B. Ivanov, ed.), Marcel Dekker, New York,

1988, pp. 379–496.

62. A. Vrij, F. T. Hesselink, J. Lucassen, and M. van den Tempel, Waves in thin

liquid films. II. Symmetrical modes in very thin films and film rupture, Proc. K.

Ned. Acad. Wet. B 73, 124–135 (1970).

63. M. Coke, P. J. Wilde, E. J. Russell, and D. C. Clark, The influence of surface

composition and molecular diffusion on the stability of foams formed from

protein/surfactant mixtures, J. Colloid Interf. Sci. 138, 489–504 (1990).

64. P. Walstra and P. E. A. Smulders, Emulsion formation, in Modern Aspects of

Emulsion Science (B. P. Binks, eds.), The Royal Society of Chemistry,

Cambridge, 1998, pp. 56–99.

65. H. Karbstein and H. Schubert, Developments in the continuous mechanical

production of oil-in-water macro-emulsions, Chem. Eng. Proc. 34, 205–211

(1995)

66. J. Chen, E. Dickinson and G. Iveson, Interfacial interactions, competitive

adsorption and emulsion stability, Food Struct. 12, 135–146 (1993).

67. H. J. Schulze, K. W. Stockelhuber, and A. Wenger, The influence of acting

forces on the rupture mechanics of wetting films–nucleation or capillary waves,

Colloids Surfaces A 192, 61–72 (2001).

68. B. V. Derjaguin and Yu. V. Gutop, Doklad. Akad. Nauk SSSR 153, 859 (1963).

69. S. Kumar, R. Kumar, and K. S. Gandhi, Influence of the wetting

characteristics of the impellor on phase inversion, Chem. Eng. Sci. 46,

2365–2367 (1991).

70. J. T. Davies, Emulsions, Academic Press, New York, 1964, pp. 129–167.

71. N. E. Hotrum, T. van Vliet, M. A. Cohen Stuart, and G. A. van Aken,

Monitoring entering and spreading of emulsion droplets at an expanding air/

water interface: a novel technique, J. Colloid Interf. Sci. 247, 125–131 (2002).

72. D. J. M. Bergink-Martens, H. J. Bos, and A. Prins, Surface dilation and fluid

dynamical behaviour of newtonian liquids in an overflowing cylinder, II.

Surfactant solutions, J. Colloid Interf. Sci. 165, 221–228 (1994).

73. C. G. J. Bisperink, The influence of spreading particles on the stability of thin

liquid films, Thesis, Wageningen University, Wageningen, The Netherlands,

1997.

74. S. Giasson, J. Israelachvili, and H. Yoshizawa, Thin film morphology and

tribology study of mayonnaise, J. Food Sci. 62, 640–646 (1997).

75. P. R. Garrett, Recent developments in the understanding of foam generation

and stability, Chem. Eng. Sci., 48, 367–392 (1993).

76. P. G. de Gennes, Two remarks on wetting and emulsions, Colloids Surfaces A

186, 7–10 (2001).

77. K. Boode and P. Walstra, Partial coalescence in oil-in-water emulsions.

1. Nature of the aggregation, Colloids Surfaces A 81, 121–137 (1993).

78. K. Boode, P. Walstra, and A. E. A. de Groot-Mostert, Partial coalescence in

oil-in-water emulsions. 2. Influence of the properties of the fat, Colloids

Surfaces A 81, 139–151 (1993).


9Orthokinetic Stability ofFood Emulsions

Siva A. VanapalliUniversity of Michigan, Ann Arbor, Michigan, U.S.A.

John N. CouplandThe Pennsylvania State University, University Park,Pennsylvania, U.S.A.

I. INTRODUCTION

Many foods and food ingredients are oil-in-water emulsions. The functionalproperties of the emulsion depend on its microstructure, which is, in turn,controlled by the ingredients and processes used. Thus, it is useful toestablish predictive links between (a) ingredients/process and microstructureand (b) microstructure and functionality. In the present context, our mainconcern is how shear, which is ubiquitous to almost all process operations,influences droplet aggregation (a microstructural feature) in emlusions.

There are very many mechanisms by which emlusions may break down,perhaps the most important are flocculation and, in the case of semicrystal-line droplets, the related phenomenon of partial coalescence. Both mecha-nisms lead to the formation of extensive networks of droplets that canincrease the viscosity of a product and lead to eventual coalescence andoiling-off (1). Flocculation can occur by a number of mechanisms andcause a wide range of changes in bulk functional properties, but in allcases, we can identify the fundamental event as two isolated dropletsbeing held at a finite separation for a significant period of time. Theremust be a minimum in the interdroplet pair potential to hold the particlestogether, but to reach this separation, the droplets must first approach one


another. It is here that applied shear forces have their greatest effect.Droplets encounter one another in a quiescent fluid through Brownianmotion, but this rate may be increased by imposed movement in the fluid.Additionally, the kinetic energy of a particle may enable it to overcome localrepulsion in the pair potential and allow droplets to aggregate, or the shearforces may rupture existing flocs. The relationships between applied flowand structure is further complicated by the fact that the structures formedand broken will, in turn, affect the emulsion viscosity and, hence, the rate ofstrain caused by an applied stress.

Despite the wide range of shear forces at all stages of food preparationand use (see Table 1) and their theoretical importance on emulsion stability,there have been very limited quantitative considerations of shear as a vari-able in the food science literature. This work is an attempt to summarize thebasic theories of particle aggregation under shear in a format useful for foodscientists and, second, to review some of the studies of the effects of shear onthe stability of emulsions. Our major focus is on floc formation and fractureand we neglect the related topic of emulsion formation (see Ref. 4). Wesimilarly neglect any treatment of the effects of emulsion structure on theshear properties of the bulk fluid [i.e., emulsion rheology (5)]. We begin byreviewing the fundamental Smoluchowski theory of colloidal aggregationand some of the more recent attempts to deal with its limitations. We nextconsider trajectory analysis, which is a complementary approach toSmoluchowski theory. Finally, we review some experimental studies onorthokinetic stability relevant to food emulsions.

II. THEORETICAL APPROACHES

The mechanism of aggregation involves two major steps: the transport step,leading to collision between two droplets, and the attachment step, in whichthe droplets stick to each other. Collisions occur due to local variations in

Table 1 Typical Shear Rates Encountered in Food Emulsions

Situation operation Shear rate (s�1)

Creaming 10�6–10�3

Pouring 10�2–102

Chewing and swallowing 101–102

Mixing and stirring 101–103

Pumping 100–103

Source: Adapted from Refs. 2 and 3.


fluid/droplet velocities arising from (a) random thermal movement of drop-lets due to Brownianmotion ( perikinetic flocculation), (b) externally imposedvelocity gradients from mixing (orthokinetic flocculation), and (c) differ-ences in creaming velocities of individual droplets (differential flocculation)(Fig. 1). In general, for particle sizes >1 mm, perikinetic aggregationbecomes less significant and orthokinetic aggregation predominates (6).For particle sizes >10 mm, differential flocculation becomes important (7).

The flux of particles through a collision surface is typically representedby the collision frequency. Because not all collisions lead to attachment,collision efficiency is used to describe the percentage of collisions that leadto aggregation. Collision efficiency is governed by the interdroplet potentialarising from the sum of colloidal and hydrodynamic interactions. The rateof droplet flocculation (J) can be taken as the product of a collision rate andcollision efficiency, so that

Jij ¼ ��ijninj ð1Þ

where � denotes the collision efficiency (whose value lies between 0 and 1)and �ij is the collision frequency between droplets of size i and j and ni and njare the number concentrations of droplets of sizes i and j, respectively. Theproduct ��ij represents the number of successful collisions leading to floccu-lation and is, therefore, sometimes referred to as flocculation frequency. Thecreation of an aggregate of a given size necessarily means the loss of one ortwo aggregates of another size, hence, the changing number in a given sizeclass can be represented by a population balance equation:

dnp

dt¼

1

2

X

iþj¼p

�ij�ijninj �X1

i¼1

�ip�ipninp ð2Þ

Figure 1 Various mechanisms of particle collision: (a) perikinetic, particles move

by random Brownian motion and collide; (b) orthokinetic, particles move in a

velocity gradient and the velocity difference between two particles leads to collision;

(c) differential, particles cream at different rates because of their different sizes and

the velocity gradient leads to a collision.


The first term on the right-hand side denotes the formation of particlesof size p due to aggregation of particles of sizes i and j (e.g., a single par-ticle, i¼ 1, and an aggregate containing five particles, j¼ 5, will form anaggregate composed of six particles, p¼ 6). The factor of 1

2 is to compensatefor counting the same collisions twice. The second term represents the loss ofparticles of size p due to collision with other particles. Equation (2) summar-izes the information we need to properly understand the kinetics of particleaggregation [i.e., np (t)], but, in practice, it is too complex to be of directvalue. For each value of p in Eq. (2), there is a differential equation to besolved (assuming �ij and �ij are known). A realistic aggregate may containseveral hundred particles, at which point the complexity is overwhelming.The useful theory arises from attempts to find a workable approximation toEq. (2).

A. Smoluchowski Theory

The first attempt to describe the flocculation process in a population ofparticles was made by von Smoluchowski (8), who started by making severalsignificant assumptions (listed in Table 2). Using these assumptions,Smoluchowski derived simple analytical expressions of collision frequencyunder perikinetic and orthokinetic conditions as shown in Eqs. (3) and (4),respectively:

�ij,perikinetic ¼2kT

3�ð3Þ

�ij,orthokinetic ¼4�

3a3 ð4Þ

where k is the Boltzmann constant, T is the temperature, � is the viscosity ofthe continuous phase, a is the droplet diameter, and � is the velocity

Table 2 Assumptions Used by von Smoluchowski in Deriving

Particle Collision Rates

1. Particles are spherical after collision.

2. All collisions lead to immediate and complete coalescence (i.e., collision

efficiency � equals unity and no colloidal or hydrodynamic forces exist).

3. There is no fracture of particles or aggregates.

4. Fluid motion is exclusively either diffusional or exclusively laminar shear.

5. Particles are of identical size.

6. Only two-particle collisions occur.

Source: Ref. 8.


gradient in the fluid. From Eq. (3), it can be seen that the collision frequencyfor perikinetic flocculation is independent of droplet size. To the contrary,for orthokinetic flocculation, Eq. (4) dictates that the collision frequencyshould increase with the cubic power of droplet size. Therefore, for finedroplets, applying a shear field may cause little change in the collisionrate, whereas for larger droplets, there will be a large increase in the rate.The total number concentration (nt) of the particles at any time t underorthokinetic conditions is given by the solution of the population balanceequation (taking into account assumptions in Table 2) as

nt ¼ n0 exp �4��t

�

� �

ð5Þ

where n0 is the initial total concentration of the particles and � is the volumefraction of the particles. Equation (5) predicts that if the assumptions of thismodel are correct, then the number of particles is predicted to decreaseexponentially with time at a rate proportional to volume fraction andshear rate.

Although Smoluchowski theory remains the basis for understandingparticle collision rates, there remain fundamental problems. Most of theassumptions listed in Table 2 are violated in real systems and so the collisionefficiency term � becomes an adjustable parameter without any theoreticalbasis. Several approaches have been taken to avoid one or more of theassumptions and, in particular, the methods of trajectory analysis haveproved as a useful complementary approach to Smoluchowski theory.

B. Assumptions of Smoluchowski Theory

Smoluchowski theory remains the basis for understanding collision rates insheared and unsheared emulsions. Several authors have made modificationsto the Smoluchowski theory to avoid some of the assumptions violated inreal systems. To some extent, it is possible to correct for these assumptionseither with an entirely empirical collision efficiency or by using a morerigorous approach such as trajectory analysis (see Section II.C).

In this subsection, we will take the approach of Thomas et al. (9) andreevaluate Smoluchowski’s assumptions and how other workers haveattempted to deal with violations.

1. Particles Are Spherical in Shape After Collision

Although liquid emulsion droplets are spherical, we are often more con-cerned with their flocculation rather than calescence and, hence, in the


formation and properties of extended flocs. The Smoluchowski derivationfor the rate of aggregation assumes that the particles remain spherical aftercollision; that is, the volumes of the two particles are additive, so that for anaggregate of N particles each of volume v and radius r, the total volume ofthe aggregate, VN, and its radius, RN, is given as

VN ¼ Nv ð6Þ

RN ¼ N1=3a ð7Þ

However, experiments on shear-induced aggregation of solid particles haveshown that the aggregates are not spherical but rather have an irregularstructure (10–12) frequently described using fractal analysis (13,14). For afractal aggregate, Eq. (7) is written as

RN ¼ N1=Da ð8Þ

where D is the mass fractal dimension. If D¼ 3 (as assumed in theSmoluchowski approach), then the structure is completely space filling. If3>D> 1, then the volume of the aggregated structure is greater than thesummed volume of the parts and so the collision radius of a fractal aggre-gate is always greater than that for an equivalent spherical aggregate. Thelower the fractal dimension, the more open (porous) the aggregate structure.If D¼ 1, the aggregated structure is linear, implying an aggregate consistingof a string of particles. Torres et al. (12) obtained a fractal dimension ofD¼ 1.8 for rapid coagulation under laminar shear flow and have shown thatthe floc structure (fractal dimension) was independent of shear rate. Inaddition, their experiments reveal that the floc structure is similar for peri-kinetic and orthokinetic aggregation, but the growth kinetics are different.Oles (11) showed that the initial growth phase of aggregation can be char-acterized by a fractal dimension of 2.1 and that later stages of growth yieldD¼ 2.5.

The presence of fractal flocs has important consequences for particlecollision frequencies and efficiencies, as both of them are functions of par-ticle size. In general, a decrease in fractal dimension enhances particle colli-sion rates (15) (due to a larger collision profile) and increases collisionefficiency (due to reduced viscous resistance to fluid flow; see Section II.C)(16). For fractal flocs, the orthokinetic collision frequency is similar tothe expression developed by Smoluchowski [Eq. (4)] but uses the fractalaggregate radius from Eq. (8):

�ij ¼4�

3a3N3=D ð9Þ


2. All Collisions Lead to Immediate and Complete Coalescence

The collision efficiency can take on a value less than 1 because of somerepulsive interaction between the approaching particles. This may be dueto colloidal forces or the hydrodynamic effects acting on the droplets.Colloidal forces are the result of the various noncovalent interactions thatoccur between particles. Each interaction decreases in magnitude in a spe-cific manner with separation, and if particles move, they must do so either inresponse to or in opposition to the sum of forces acting at a given separa-tion. This can be mapped as an interaction pair potential which describes theenergy cost to move one particle from infinite separation to a given distancefrom a second, fixed particle and results from a summation of thevarious forces acting. One of the most commonly used pair potential isthe DVLO potential (given as a sum of electrostatic repulsive and van derWaals attractive forces) and this is illustrated in Fig. 2.

At long separations, the interaction between the particles is effectivelyzero, but as they approach closer, the repulsive electrostatic interactionsdominate over the attractive van der Waals forces and the particles arerepulsed. At very close separations, the attractive forces dominate and theparticles aggregate. (Note that for hard particles, there will be a very strongshort-range steric repulsive force added to prevent coalescence.) The prob-ability of two particles on a collision path approaching one another to thepoint of impact then depends on their kinetic energy and the potentialenergy surface over which they move. The important features of the inter-droplet potential are the depth of the close separation energy well and the

Figure 2 Diagrammatic representation of the DVLO interaction potential.


height of the intermediate separation barrier. Knowing these, it is possibleto calculate a value for � under perikinetic conditions (17).

Hydrodynamic forces represent the energy required to make the fluidbetween the approaching particles to ‘‘get out of the way.’’ They will actagainst particle approach and are thus repulsive forces. The relative move-ment of the fluid will also tend to cause approaching particles to rotate andtake on a curved trajectory and miss one another. (Trajectory analysis iscovered in more depth in Section II.C). Consequently, the presence ofhydrodynamic forces will always reduce the collision frequency, sometimesby up to five orders of magnitude, compared to rectilinear trajectories (18).It is possible to calculate the effects numerically (7,18) and it is notable thatthe most dramatic reduction is in the case of orthokinetic flocculation—particularly with a wide disparity in particle sizes.

The structure of the flocs can affect the colloidal forces actingbetween them:

. Impermeable porous flocs. In the case of impermeable porous flocs(i.e., no passage of fluid through the floc), the collision efficiency(obtained from trajectory analysis) can be corrected for the factthat the van der Waals force between porous flocs is smaller thanthat of solid particles. This is so because the attractive forcebetween two flocs can be considered to be that between the nearestprimary particles, because the other particles in the floc areseparated by too large a distance (19). Using this hypothesis,Kusters et al. (16) showed that collision efficiency betweenimpermeable flocs decreases more rapidly with increasing particlesize than that for solid particles.

. Permeable porous flocs. For permeable porous flocs, the hydro-dynamic forces are less pronounced, due to penetration of fluidflow, resulting in higher values of collision efficiencies. The Debyeshielding ratio, �, defined as

� ¼Rffiffiffi�

p ð10Þ

denotes the extent of passage of fluid flow, where R is the outerradius of floc and � is the floc permeability. Porous flocs can also berepresented using a shell–core model, which has an impermeablecore with a completely permeable outer shell. The outer collisionradius represents the distance within which another floc mustapproach for aggregation to occur. The inner impermeable coreradius describes the drag experienced by the floc and corres-ponds to the hydrodynamic radius (RH). With the knowledge of


hydrodynamic functions for hard spheres (20), the trajectories of thetwo colliding impermeable flocs are calculated and the outer radiusis used to determine the point of contact. In this way, the influenceof fluid flow penetration is accommodated in the collision efficien-cies that follow from trajectory analysis (see Section II. C).Colloidal forces have not yet been satisfactorily incorporated inthis model of floc structure. Nevertheless, the existing shell–coremodel has been shown to be in good qualitative agreement withthe experimental results for aggregate growth under turbulent flowconditions (12,21).

More recently, experiments have demonstrated a nonuniform internalstructure of the floc, in contrast to the uniform porosity assumed in theshell–core model (22). Collision efficiency of fractal flocs can be predicted byperforming trajectory analysis, using the shell–core model described earlier(16). The floc structure is incorporated into the shell–core model using theDebye-shielding ratio that depends on the outer collision radius (which is afunction of fractal dimension).

3. There Is No Fracture of Particles or Aggregates

Although neglected in Smoluchowski theory, in most systems flocs canbreak under the applied shear forces and the actual size distribution atany given time must account for both formation and destruction rates.The breakup of flocs occurs at a critical shear rate, at which the hydrody-namic force exceeds the attractive forces holding the flocs together; that is,

3

2��R2� FA ð11Þ

where R is the floc radius and FA is the attractive force between the flocs(12). Two distinct modes of floc breakage have been reported in the litera-ture under turbulent conditions: floc erosion and floc fracture (23). Flocerosion corresponds to the removal of primary particle from the surfaceof the floc and, in turbulent flow, predominates when floc size is smallerthan the Kolmogorov scale. Floc fracture occurs in larger flocs and is due tothe fluctuating motion of the fluid. Yeung and Pelton (24) showed experi-mentally that compact flocs are more susceptible to undergo surface erosionand less compact flocs are likely to fracture. Using a mean-field approach,Sonntag and Russel (25) developed a model for breakup that incorporatesfloc density in the form of fractal dimension. Their model was able topredict the shear-rate dependence of floc size, although their modelneglected the rearrangement of flocs after breakup.


In practice, a flocculating system of particles, subjected to shear willeventually reach a steady state where the rate of floc fracture matches therate of floc formation. The shear-rate dependence on the maximum stablefloc size (dF,max) is given by (26,27):

dF , max ¼ b��m ð12Þ

where b and m are constants determined by the strength of the attractiveforces and the internal floc structure, respectively. However, experimentalresults show that the effects of shear on maximum floc size could be quitecomplex. For example, restructuring of aggregates was observed at moder-ate shear rates (40–80 s�1) but not at lower shear rates (28). Further, longershearing times could lead to the formation of compact flocs (with higherfractal dimensions) due to restructuring (11,29). If the particle concentrationis below a certain critical value, the maximum aggregate size dependsonly on shear rate, otherwise it is a function of both shear rate andconcentration (30).

4. Fluid Motion Is Exclusively Laminar Shear

The approach taken by Smoluchowski was essentially a two-dimensionalsimplification of three-dimensional laminar flow. In most realistic cases,flow is three dimensional and it is more complex to define the motion ofone particle relative to another. Camp and Stein (31) developed an alter-native version of Eq. (4), replacing the velocity gradient with the root-mean-square velocity gradient (�rms), and although this approach is flawed (9), itretains considerable practical value. Root-mean-square velocity was relatedto the local rate of energy dissipation, ", in laminar flow and the kinematicviscosity (�) as

�rms ¼

ffiffiffi"

�

r

ð13Þ

Camp and Stein further extended this idea to turbulent flow by defining abulk root-mean-square velocity (��rms) in terms of the bulk rate of energydissipation ("�). Incorporating the fractal dimensionality of realistic flocs,the collision rate is given as [see Eqs. (4) and (9)]

�ij ¼ 10:352"�

�

� �1=2

a3N3=D ð14Þ

5. Particles Are of Identical Size

Although fine liquid droplets are necessarily spherical, real emulsions arerarely monodisperse. Even for a hypothetical monodisperse emulsion,


Smoluchowski’s assumption would soon be violated as doublets and tripletsformed early in the aggregation process start to take part in further reac-tions. Solving for all of the rate constants in Eq. (2) is impractical for a largerange of sizes; some authors have made assumptions instead about theallowable particle sizes:

. The allowable particle size must follow a predefined mathematicalseries (e.g., geometric, n¼ 1, 2, 4, 8, 16,. . .). The loss of sensitivity isthe cost for the reduced complexity of the calculations.

. The initial particle size distribution displays properties of self-preservation; that is, as the system aggregates, the entire distribu-tion shifts to larger sizes but retains the same shape. In suchinstances, a correction factor can be introduced to collisionefficiency to account for polydispersity (32).

6. Only Two-Particle Collisions Occur

As systems become progressively more concentrated, this assumptionbecomes less valid; however, the theoretical demands of multibody collisionshave not yet been adequately incorporated into Smoluchowski theory.

Despite its limitations, Smoluchowski theory, particularly as modifiedby Camp and Stein [Eq. (13)], remains the basis for understanding dropletcollision behavior. However, as we have seen, it has many limitationsrestricting its applicability and many of the approaches to deal with itsassumptions are largely empirical. A more detailed and fundamentalapproach can be gained through trajectory analysis.

C. Trajectory Analysis

The limitations of Smoluchowski theory are typically addressed through theadjustable parameter �, the collision efficiency. This approach is broadlymacroscopic and allows no detailed description of the nature of the inter-actions taking place. If we know all of the forces involved, it would betheoretically possible to simulate the aggregation process through computersimulation (Monte Carlo or colloidal dynamics). However, it is often pre-ferred to consider only two-body interactions in more depth. Trajectoryanalysis is a powerful, although somewhat complex, tool to understandthe motion of one particle relative to another, but it is only suitable fortwo-body interactions and, thus, is only quantitatively valid for very dilutesuspensions ð� 0:01Þ. We shall introduce some of the important conclu-sions from trajectory analysis in the following paragraphs, but for a more in


depth discussion, the reader is advised to refer to Van de Ven’s book (33)and a recent review by Vanni and Baldi (34).

Trajectory analysis is a method to determine the path of interactionof two spheres in a fluid flow. It is assumed that one of the spheres is fixedat the origin (reference sphere) and the relative trajectory of the othersphere (collision sphere) is then calculated (Fig. 3). Another key assump-tion is that the interaction forces act along the line joining the centers ofthe two spheres and are balanced by the component of the hydro-dynamic force in that direction. The equations governing the trajectoriesof the two particles in a simple laminar shear flow are given by Vanni andBaldi (34):

dr

dt¼ �rð1� AÞ sin2 � sin� cos�þ

CEint

6��a1ð15Þ

d�

dt¼ �ð1� BÞ sin � cos � sin� cos� ð16Þ

d�

dt¼ � cos2 ��

B

2cos 2�

� �

ð17Þ

where � is the shear rate and the angles � and �, and the distance r representthe relative positions of the particles in radial coordinates (see Fig. 3). Fint isthe interparticle pair potential, and A, B, and C are hydrodynamic mobilityfunctions which depend on the dimensionless center distance r� ¼ r/a1 andon the size ratio ¼ a2/a where a is the particle radius and subscripts 1 and 2refer to the reference and collision spheres, respectively. Approximate

Figure 3 Diagram showing the axis system used for trajectory analysis of the

interaction between a reference sphere (1) and a collision sphere (2). The flow

direction is along y axis and the velocity gradient is along the x axis.


expressions for the hydrodynamic (34) and the interparticle pair potentialfunctions (35) are available in the literature.

The trajectory of the collision sphere is determined by numericallyintegrating Eqs. (15–17) (Fig. 4a). The numerical integration requiresknowledge of the initial position and velocity of the particle. The collisionsphere is made to approach the reference sphere from an infinite distance(at least 10 particle separations) and can have 3 possible outcomes. Itcan undergo collision (i.e., primary coagulation: r/a1<1þ a2/a1þ ",

Figure 4 (a) Limiting trajectories of the interaction of a collision sphere (2) with a

reference sphere (1). The broken circle indicates the region through which sphere 2

must pass to collide with sphere 1. The rectilinear trajectory starts at the minimum

distance from the origin for sphere 2 to miss sphere 1 using Smoluchowski

assumptions (i.e., rectilinear trajectories). The curvilinear trajectory starts at the

minimum distance from the origin for sphere 2 to miss sphere 1 when hydrodynamic

forces are included (i.e., curvilinear trajectories). Note that yc<a1þ a2 when

hydrodynamic forces act, so fewer particles in the flux will collide. Again, the y-x

plane is the flow-gradient plane. (b) Collision surfaces calculated from trajectory

analysis. The surface marked rectilinear is that predicted from the Smoluchowski

(i.e., no colloidal forces acting). The surface marked curvilinear is a sample result

that may occur when colloidal and hydrodynamic forces are taken into account.

(Adapted from Ref. 34.)

(a)


where " is a predetermined parameter set in the numerical integration)or orbit around the reference sphere (i.e., secondary coagulation: � > �=2)or simply depart to infinity without contacting the reference sphere( y/a1>10). The calculation is repeated from a number of starting positionsand a collision surface is mapped.

A collision surface can be calculated as those starting positions forfluid flow that would result in particle collision (Fig. 4b). If interparticleforces and hydrodynamic forces are neglected, then the collision spherewould follow the stream lines of the simple shear flow (i.e., rectilinear tra-jectories). In such as case, a collision occurs for all particles that movetoward the reference particle within the cylinder defined by a1þ a2. If thedistance of approach is greater than this distance, the particles will miss eachother. The collision surface in Smoluchowski theory is a circle, whose radiusis defined by the distance of limiting trajectory from the origin (i.e., a1þ a2)(Fig. 4b).

The flux is given by

J ¼

ZdJ ¼

Z

s

NvZðsÞ ds ð18Þ

where N is the number of particles crossing the collision surface Z(s) and vrepresents the particle velocity. Taking Z(s) as a circle of radius a1þ a2,simple integration yields the Smoluchowski flux equation:

J0 ¼4

3�Nða1 þ a2Þ

3ð19Þ

[Note Eq. (19) is a general case of Eq. (4) for dissimilar particle sizes.] Whenonly hydrodynamic forces are acting, then the trajectories become curvi-linear (Fig. 4a) and the definition of the collision surface is no longerobvious (Fig. 4b). The hydrodynamic forces cause the collision particle toroll past the reference particle and miss even though it started within thea1þ a2 limit for collision if the path had been rectilinear (Fig. 4a;yc< a1þ a2). Note that the trajectory is symmetric about the y axis; that is,the positions of the collision particle at the upstream and downstreamregion are mirror images of each other. The full problem of incorporatingcolloidal and hydrodynamic forces into trajectory analysis has been solvedby various authors (36–39) to calculate the collision surface. The collisionsurface in this case could be quite irregular, as shown in Fig. 4b. In this case,the flux is given by

J ¼ 4�N

Z zmax

0

xzðxÞ dx ð20Þ


Note that in this case the trajectories are asymmetrical about the y axis,because the collision sphere spends a much longer time in the vicinity of thereference sphere due to the colloidal interaction force. In addition, the curvi-linear trajectory in Fig. 4b extends beyond the Smoluchowski limit ofa1þ a2, suggesting that some of the particles may stick in the downstreamregion after crossing the x-z plane.

The collision efficiency is then given by the ratio of Eqs. (19) and (20)(i.e., the ratio between actual collision rates incorporating all forces and theideal collision rate with no forces acting):

� ¼3

ða1 þ a2Þ3

Z zmax

0

xzðxÞ dx ð21Þ

By this method, it is possible to directly calculate the adjustable parameter,�, in Smoluchowski theory from known physical properties of the system.This formulation of trajectory analysis assumes diffusional mobility of theparticles is insignificant compared to shear. For particle aggregation undershear flow, Swift and Friedlander (40) showed that the Brownian and bulkconvective motion could be treated independent of each other and thusmerely added to determine the combined effect; however van de Ven andMason (41) showed that the additive assumption is incorrect whenBrownian effects were large compared to shear forces [i.e., at very low Pe(the Peclet number)].

This analysis is strictly for solid droplets and there are some importantdifferences for fluid particles, which can deform due to the forces acting.The hydrodynamic functions for interacting droplets depend not only on thesize ratio of the two droplets and the dimensionless interdroplet distance butalso on the droplet–medium viscosity ratio because the particles can deformsomewhat to absorb the hydrodynamic stress (42,43).

III. EXPERIMENTS IN FOOD SYSTEMS

Finally, we shall consider some of the experimental studies on food emulsionstability that have considered shear as a variable. Most of the orthokineticstudies in food emulsions have been done in fairly concentrated emul-sions (10–30wt%) and will be reviewed first. Later, we will examinerecent work by the Leeds group that has developed a particle capture tech-nique to study interactions between two colloidal particles in a laminarshear field (44).


A. Orthokinetic Studies on Concentrated Emulsions

In experimental studies on shear-induced aggregation of concentrated foodemulsions, the measured parameter almost always has been the change indroplet size distribution. The variation in droplet size distributions due tocoalescence/flocculation is transformed into a coalescence rate constant orcoalescence efficiency, to quantify the effect of shear. In some cases, viscos-ity is used as a proxy for particle aggregation. Typical results show a sig-moidal function for particle size in time for emulsion destabilization undershear (Fig. 5). In the first portion of the graph, there is no change becausethe primary droplets have a very low reactivity. However, once a few drop-lets do aggregate, then the rate of reaction increases with particle size (par-ticularly important for fractal flocs). However, large flocs are eroded (orfractured) by the velocity gradient, and as the size increases, this process willmatch the rate of aggregation and reach an equilibrium value of particlesize. This equilibrium may change if the shear rate is changed or stopped.

1. Emulsions Containing Liquid Droplets

Van Boekel and Walstra studied the effect of laminar Couette flow on thestability of paraffin oil-in-water emulsions stabilized by various surfactants.The experimental rate constant for coalescence (kexp) was taken as the slopeof a linear fit to a plot of droplet size with time (on a semilog scale). Thetheoretical rate constant of coalescence was derived from Smoluchowskirate equation

dnt

dt¼ �ktheorynt ð22Þ

Figure 5 Schematic diagram showing typical changes in particle size for an

emulsion undergoing orthokinetic flocculation.


which is obtained by differentiating Eq. 5, where

ktheory ¼4�’

�ð23Þ

They quantified the effect of shear in terms of coalescence efficiencydefined as the ratio of kexp to ktheory. The orthokinetic studies showed thatemulsions that were stable at rest were also stable during Couette flow(45,46). In addition, they observed that in emulsions that were unstable atrest, Couette flow did not have any significant influence on the rate ofcoalescence. Increasing the shear rate would increase the frequency of colli-sions, but these authors argued that it would simultaneously decrease theduration of encounters (i.e., collision efficiency decreased with shear rate inCouette flow). Attempts were made by these authors to increase the ortho-kinetic collision efficiency through the addition of NaCl, but the emulsionsremained resistant to shear-induced coalescence. This led them to suggestthat the duration of an encounter is more important than the number ofencounters per se. In addition, the authors performed trajectory analysis tocalculate the collision efficiency for doublet formation, which was muchhigher than the measured value.

In contrast, under turbulent flow, a dramatic increase in aggregationrate after a few larger, more reactive particles were formed was seen byDickinson and Williams (47) in a study of the orthokinetic stability ofprotein-stabilized emulsions. They observed that there was a sharp diver-gence in d32 after an initial lag phase and defined a characteristic destabili-zation time at the maximum rate of increase in the mean droplet diameter.They proposed the sudden destabilization in terms of an increasing reactiv-ity of droplets with increasing droplet size (� is a strong function of particlediameter), similar to the critical phenomenon in a sol–gel transition (48) andused a two-parameter model to describe the process:

dnt

dt¼ kdnt � kcn

2t ð24Þ

where kd and kc are the rate constants for disruption and coalescence,respectively. The coalescence rate constant was assumed to depend on drop-let size and fractal dimension [see Eq. (9)]. They further invoked a minimumdroplet size breakup criterion [see Eq. (11)]. This model showed that afractal dimension D� 2 was required to match the experimental data,which is in reasonable agreement with other orthokinetic studies. Theseand other workers argued that attractive colloidal forces make emulsionsmore vulnerable to shear-induced aggregation. For example, adjusting the


pH toward the pI of the protein-stabilizing layer (47) and adding calcium toa caseinate-stabilized system (49,50) have all been shown to increase thesusceptibility to orthokinetic destabilization. Furthermore, adding a smallamount of small-molecule surfactant to a protein-stabilized emulsionincreased the orthokinetic destabilization rate without causing significantdisplacement (51,52). (It was argued that the increased membrane fluiditycaused by adsorbed surfactant was responsible.) On the other hand,increased amounts of interfacial protein (presumably forming a thickerlayer) decreased the susceptibility to orthokinetic destabilization (50).

As well as increasing the rate of orthokinetic aggregation, increasedshear rates can limit the maximum (equilibrium particle size) achieved byincreasing the rate of aggregate fracture. For example, Schokker andDalgleish (50) showed that reciprocal maximum particle diameter was pro-portional to the shear rate used for a caseinate-stabilized emulsion [i.e.,m¼� 1 in Eq. (11)]. Again, by increasing the magnitude of the attractiveforces between aggregated particles, which oppose the shear-induced stresseson the floc [e.g., by adding calcium to a caseinate stabilized emulsion(49,50), the final floc size achieved can be increased].

2. Emulsions Containing Semicrystalline Droplets

For liquid-oil emulsions, aggregation is generally due to interactionsbetween the surface layers and the oil composition itself has limited effect.However, in many emulsions of practical interest, the oil phase may bepartially crystalline at usage temperatures (46). The crystals can form asspikes at the droplet interface and the crystalline droplets are more reactivetoward other semisolid droplets (53). The fat crystals in one droplet induceinstability by penetrating the thin film between two approaching dropletsand allowing liquid oil to flow out around the interaction site to reinforcethe link. This phenomenon is known as partial coalescence because althoughthere is mixing of the oil from two droplets, the crystal network is sufficientto maintain the doublet shape. Completely solid or liquids cannot partiallycoalesce, as they lack the required liquid or solid fat fraction, respectively.Because partial coalescence requires direct droplet–droplet contact, it wouldbe reasonably expected to be highly dependent on shear rate. Walstra (1)argued that the applied shear (a) increases the collision rate, (b) pushes thecolliding droplets closer together so the protruding crystal size approachesthe film thickness, and (c) the hydrodynamic forces between the dropletscauses the droplets to roll around one another, increasing the probabilitythat a suitable protruding crystal will be able to penetrate the interdropletmembrane.


Partial coalescence is particularly important in milk fat emulsionsbecause butter oil has a wide melting range and under storage conditions(� 5�C) is semisolid. Milk fat globules are quite stable to coalescence whenat rest (except slow creaming). However, it has been shown that liquid milkfat globules are unstable under Couette flow (46), probably because theincreased collision rate for the large droplets is sufficient to enhancethe rate of partial coalescence. Van Boekel and Walstra (46) observed thatthe coalescence efficiency remained constant with increasing laminar shearrate, indicating that the rate of coalescence was proportional to the encoun-ter frequency and not the product of frequency and duration, as suggestedfor liquid-oil emulsions (see above).

Minor components in the fat are very capable of affecting the orienta-tion of crystals to the surface and hence the orthokinetic stability of thesemicrystalline emulsion. Davies and others (54) studied the shear stabilityof partly crystalline triglyceride droplets stabilized with sodium caseinate.The emulsions were quiescently stable but had clear stress limits that, ifexceeded, would quickly lead to emulsion destabilization. The magnitudeof these stress limits decreased with increasing solid fat content andwith increasing added monoglycerides. The monoglycerides were believedto both weaken the protein coat of the droplets and act as a templatefavoring crystal growth/accumulation at the surface (55). Hinrichs andKessler (56) also noted the presence of a critical limit in the strain raterequired to destabilize milk fat globules. The magnitude of this criticallimit decreased linearly with fat volume fraction and increased with solidfat content.

B. Two-Particle Aggregation Under Shear

The macroscopic approaches used in the above-described studies, althoughpractical, does not provide quantitative information regarding the interplayof colloidal and hydrodynamic forces. Recently, the role of such forces infood systems was studied by the Leeds Group using a colloidal particlescattering technique (57–60). In this technique, the capture of one dropletmoving in a shear field by a second (fixed) droplet was observed. Thismicroscopic approach yields trajectories of the moving droplet, whichwere then compared with computer simulations of the trajectories (similarto the trajectory analysis discussed in Section II. C) to obtain quantitativeinformation about dynamic colloidal interactions. In addition, the captureefficiency (which is qualitatively similar to collision efficiency described inSection III. A) was determined, which was defined as the ratio of numberof collisions leading to the formation of a doublet to the total number oftrials.


The particles used in the particle-scattering technique were latex andemulsion droplets, coated with protein emulsifiers including sodium case-inate, �-casein, and �s1-casein. They observed that in caseinate-stabilizedemulsions, the capture efficiency was highest at the isoelectric point(pH¼ 4.7) (60), as expected from the macroscopic studies (e.g., Ref. 47).To achieve a good fit with the experimental data, it was necessary to developa theory for the colloidal forces acting between the droplets [see Eq. (15)].These workers initially used a DLVO potential (as shown in Fig. 2) with anadditional DeGennes steric repulsive interaction but this agreed poorly withthe data. Further, this model did not predict the observed upstream stickingof the mobile particle to the fixed particle. Only by incorporating a pH-dependent tangential interparticle force, which they hypothesized couldarise from entanglement of the adsorbed polymer layers at the interface,were they able to successfully predict the variation of capture efficiency withpH. When ionic strength was included as an additional parameter, thebehavior was complex. At lower pH (<5.2), sticking probability was sensi-tive to lower values of ionic strengths (<0.2M ), indicating the importanceof charged interactions in shear-induced aggregation. At higher ionicstrength (¼0.5M ), it was unclear as to why the capture efficiency woulddecrease dramatically.

IV. CONCLUSIONS

We are interested in the changing structure and functionality of food emul-sions over long periods of quiescent storage and short periods of varyingflow during manufacture and use. The forces acting on the droplets in eachcase and so the behavior of the emulsions can be very different. Althoughthe basic Smoluchowski approach is often invalid in sheared systemsbecause of the strong effect of hydrodynamic forces, it is possible to usemodified theories, particularly in combination with trajectory analysis, togive good predictions of collision rates in dilute emulsions. However, asconcentration is increased, the theoretical assumptions are violated andthe theories may prove unreliable. Furthermore, the colloidal interactionpotential needed for trajectory analysis are difficult to know precisely apriori for a realistic system and can easily become a hidden fitting para-meter. The particle scattering approach taken by Dickinson and others(44,57–60) provides a more experimental approach to verify the behaviorof pairs of particles, although the possibility of multibody interactions athigher concentrations cannot be considered.

Experimental study of emulsion destabilization in flow is also chal-lenging. Modern rheometers are ideal for providing controlled flow rates


in a fluid, but the characterization of the response of the suspension ismore complex. Increased viscosity is a sensitive measure of droplet floc-culation (but not to coalescence) and is relatively easy to measure in situif the rheometer is also used to generate the fluid flow (e.g., Ref. 54).However, although it is possible to infer aggregate structure from rheo-logical measurements, this is an indirect approach. Removing samplesfrom the flow for analysis by conventional light scattering (48) or micro-scopy (61–64) is possible, but this has the disadvantage of changing thebalance of forces acting and possibly the structures present (65). Someauthors have combined rheometers and light-scattering instruments tostudy aggregation under shear, but this is necessarily restricted to dilutesystems (12,28,66). Other methods of particle sizing suited to concentratedemulsions, particularly ultrasonic spectroscopy, may be more suited to thestudy of suspension structure under shear (61,64). Hamberg and others(67) recently used several methods to study the changes in the structure oflatex particles covered with whey protein in shear flow. First, by usingfluid gelatine solution as the continuous phase, they were able to easily fixsamples removed from the rheometer for optical microscopy. Secondthey used a four-roll mixer to generate a region of laminar shear flowbetween the rollers and imaged this flow using a confocal laser scanningmicroscope.

In conclusion, we can expect that the fluid flow to drastically effectemulsion structure and that studies on quiescent systems need notapply when the mixture is sheared. The effects can be both complex andunexpected.

REFERENCES

1. P. Walstra, Emulsion stability, in Encyclopedia of Emulsion Technology

(P. Becher, ed.), Marcel Dekker, New York, 1996, Chap. 1.

2. H. A. Barnes, Rheology of emulsions—a review, Colloids Surfaces A 91, 89–95

(1994).

3. J. F. Steffe, Rheological Methods in Food Process Engineering, 2nd ed., Freeman

Press, East Lansing, MI, 1996.

4. P. Walstra, Principles of emulsion formation, Chem. Eng. Sci. 48, 333–349

(1993).

5. E. Dickinson, Proteins at interfaces and in emulsions. Stability, rheology and

interactions, J. Chem. Soc. Faraday Trans. 94, 1657–1667 (1998).

6. V. G. Levich, Physicochemical Hydrodynamics, Prentice-Hall, Englewood Cliffs,

NJ, 1962.

7. M. Y. Han and D. F. Lawler, The (relative) insignificance of g in flocculation,

J. Am. Water Works Assoc. 84, 79–91 (1992).


8. M. Smoluchowski, Verusch einer Mathematischen Theorie der

Koagulationskinetik Kolloider Losungen, Z. Phys. Chem. 92, 129–168 (1917).

9. D. N. Thomas, S. J. Judd, and N. Fawcett, Flocculation modeling: A review,

Water Res. 33, 1579–1592 (1999).

10. R. C. Klimpe and R. Hogg, Effects of flocculation conditions on agglomerate

structure, J. Colloid Interf. Sci. 113, 121–131 (1986).

11. V. Oles, Shear-induced aggregation and breakup of polystyrene latex particles,


12. F. E. Torres, W. B. Russel, and W. R. Schowalter, Floc structure and growth

kinetics for rapid shear coagulation of polystyrene colloids, J. Colloid Interf.

Sci. 142, 554–574 (1991).

13. B. B. Mandelbrot, The Fractal Geometry of Nature, W. H. Freeman, San

Francisco, 1982.

14. P. Meakin, Fractal aggregates, Adv. Colloid Interf. Sci. 28, 249–331 (1988).

15. Q. Jiang and B. E. Logan, Fractal dimensions of aggregates determined from

steady state size distributions, Environ. Sci. Technol. 25, 2031 (1991).

16. K. A. Kusters, J. G. Wijers, and D. Thoenes, Aggregation kinetics of small

particles in agitated vessels, Chem. Eng. Sci. 52, 107–121 (1997).

17. N. Fuchs. Uber die stabilitat und aufladung der aerosole, Z. Phys. 89, 736–743

(1934).

18. D. F. Lawler, Physical aspects of flocculation: From microscale to macroscale,

Water Res. 27, 164–180 (1993).

19. B. A. Firth and R. J. Hunter, Flow properties of coagulated colloidal

suspensions. I. Energy dissipation in the flow units, Colloid Interf. Sci. 57,

248–256 (1976).

20. P. M. Adler, Heterocoagulation in shear flow, J. Colloid Interf. Sci. 83, 106–115

(1981).

21. K. A. Kusters, The influence of turbulence on aggregation of small particles in

agitated vessel, Ph. D. thesis, Eindhoven University of Technology, Eindhoven,

1991.

22. R. C. Sonntag and W. B. Russel, Structure and breakup of flocs sub-

jected to fluid stresses 1. Shear experiments, J. Colloid Interf. Sci 113, 399–413

(1986).

23. K. Muhle, Floc stability in laminar and turbulent flow, in Coagulation and

Flocculation: Theory and Applications (B. Dobias, ed.), Marcel Dekker, New

York, 1993, pp. 355–390.

24. A. K. C. Yeung and R. Pelton, Micromechanics: A new approach to

studying the strength and break-up of flocs, J. Colloid Interf. Sci. 184,

579–585 (1996).

25. R. C. Sonntag and W. B. Russel, Structure and breakup of flocs subjected to

fluid stresses 2. Theory, J. Colloid Interf. Sci. 115, 378–389 (1986).

26. J. C. Flesch, P. T. Spicer, and S. E. Pratsinis, Laminar and turbulent shear-

induced flocculation of fractal aggregates, AIChE J. 45, 1114–1124 (1999).

27. A. A. Potanin, On the mechanism of aggregation in the shear flow suspensions,

J. Colloid Interf. Sci. 145, 140 (1991).


28. C. Selomulya, R. Amal, G. Bushell, and T. D. Waite, Evidence of shear rate

dependence on restructuring and break-up of latex aggregates, J. Colloid Interf.

Sci. 236, 67–77 (2001).

29. P. T. Spicer and S. E. Pratsinis, Shear-induced flocculations: The evolution of

floc structure and the shape of the size distribution at steady state, Water Res.

30, 1049–1056 (1996).

30. T. Serra and X. Casamitjana, Effect of shear and volume fraction on the

aggregation and breakup of particles, AIChE J. 44, 1724–1730 (1998).

31. T. R. Camp and P. C. Stein, Velocity gradients and internal work in fluid

motion, J. Boston Soc. Civil Eng. 30, 219–237 (1943).

32. V. Mishra, S. M. Kresta, and J. H. Masliyah, Self-preservation of the drop size

distribution function and variation in the stability ratio for rapid coalescence of

a polydisperse emulsion in a simple shear field, J. Colloid Interf. Sci. 197, 57–67

(1998).

33. T. G. M. Van der Ven, Colloidal Hydrodynamics, Academic Press, London,

1989.

34. M. Vanni and G. Baldi, Coagulation efficiency of colloidal particles in shear

flow, Adv. Colloid Interf. Sci. 97, 151–177 (2002).

35. J. N. Israelachvili, Intermolecular and Surface Forces, Academic Press, London,

1992.

36. D. L. Feke and W. R. Schowalter, The effect of Brownian diffusion on

shear-induced coagulation of colloidal dispersions, J. Fluid Mech. 133, 17

(1983).

37. D. L. Feke and W. R. Schowalter, The influence of Brownian diffusion on

binary flow-induced collision rates in colloidal dispersions, J. Colloid Interf.

Sci. 106, 203 (1985).

38. T. G. M. Van der Ven and S. G. Mason, The microrheology of colloidal

dispersions. VII. Orthokinetic doublet formation of spheres, Colloid Polym.

Sci. 255, 468–479 (1977).

39. G. R. Zeichner and W. R. Schowalter, Use of trajectory analysis to study

stability of colloidal dispersions, AIChE J. 23, 243–254 (1977).

40. D. L. Swift and S. K. Friedlander, The coagulation of hydrosols by Brownian

motion and laminar shear flow, J. Colloid Sci. 19, 621–647 (1964).

41. T. G. M. Van der Ven and S. G. Mason, The microrheology of colloidal

dispersions. VIII. Effect of shear on perikinetic doublet formation, Colloid

Polym. Sci. 255, 794 (1977).

42. H. Wang, A. Z. Zinchenko and R. H. Davis, The collision rate of small drops in

linear flow fields, J. Fluid Mech. 265, 161–188 (1994).

43. A. Z. Zinchenko and R. H. Davis, Collision rates of spherical drops of particles

in a shear flow at arbitrary Peclet numbers, Phys. Fluid 7, 2310–2327 (1995).

44. B. S. Murray, E. Dickinson, J. M. McCarney, P. V. Nelson, M. Whittle,

Observation of the dynamic colloidal interaction forces between casein-coated

latex particles, Langmuir 14, 3466–3469 (1988).

45. M. A. J. S. van Boekel and P. Walstra, Effect of Couette flow on stability of

oil-in-water emulsions, Colloids Surfaces 3, 99–107 (1981).


46. M. A. J. S. van Boekel and P. Walstra, Stability of oil-in-water emulsions with

crystals in the disperse phase, Colloids Surfaces 3, 109–118 (1981).

47. E. Dickinson and A. Williams, Orthokinetic coalescence of protein-stabilized

emulsions, Colloids Surfaces A 88, 317–326 (1994).

48. A. Lips, T. Weatbury, P. M. Hart, I. D. Evans, and I. J. Campbell, On the

physics of shear-induced aggregation in concentrated food emulsions, in Food

Colloids and Polymers: Stability and Mechanical Properties (P. Walstra, ed.),

Royal Society of Chemistry, Cambridge, 1993, pp. 31–44.

49. E. Dickinson and E. Davies, Influence of ionic calcium on stability of sodium

caseinate emulsions, Colloids Surfaces B 12, 203–212 (1999).

50. E. P. Schokker and D. G. Dalgleish, Orthokinetic flocculation of caseinate-

stabilized emulsions: Influence of calcium concentration, shear rate and protein

content, J. Agric. Food Chem. 48, 198–203 (2000).

51. E. Dickinson, G. Iverson, and J. Chen, Interfacial interactions, competitive

adsorption and emulsion stability, Food Struct. 12, 135–146 (1993).

52. E. Dickinson, R. K. Owusu, and A. Williams, Orthokinetic destabilization of a

protein stabilized emulsion by a water soluble surfactant, J. Chem. Soc.

Faraday Trans. 89, 865 (1993).

53. D. Rousseau, Fat crystals and emulsion stability—a review, Food Res. Int. 33,

3–14 (2000).

54. E. Davies, E. Dickinson, and R. D. Bee, Shear stability of sodium caseinate

emulsions containing monglyceride and triglyceride crystals, Food

Hydrocolloids 14, 145–153 (2000).

55. E. Davies, E. Dickinson, and R. D. Bee, Orthokinetic destabilization of

emulsions by saturated and unsaturated monoglycerides, Int. Dairy J. 11,

827–836 (2001).

56. J. Hinrichs and H. G. Kessler, Fat content of milk and cream and effects on fat

globule stability, J. Food Sci. 62, 992–995 (1997).

57. H. Casanova, J. Chen, E. Dickinson, B. S. Murray, P. V. Nelson, and

M. Whittle, Dynamic colloidal interactions between protein-stabilized

particles—experiment and simulation, Phys. Chem. Chem. Phys. 2, 3861–3869

(2000).

58. E. Dickinson, B. S. Murray, M. Whittle, and J. Chen, Dynamic interactions

between adsorbed protein layers from colloidal particle scattering in shear flow,

Food Colloids: Fundamentals of Formulation (R. Miller, ed.), Royal Society of


59. M. G. Semenova, J. Chen, E. Dickinson, B. S. Murray, and M. Whittle,

Sticking of protein-coated particles in a shear field, Colloid Surfaces B 22, 237–

244 (2001).

60. M. Whittle, B. S. Murray, J. Chen, and E. Dickinson, Simulation and

experiments on colloidal particle capture in a shear field, Langmuir 16,

9784–9791 (2000).

61. R. Chanamai, N. Hermann, and D. J. McClements, Ultrasonic spectroscopy

study of flocculation and shear-induced floc disruption in oil-in-water

emulsion, J. Colloid Interf. Sci. 204, 268–276 (1998).


62. H. Mousa, W. Agterof, and J. Mellema, Experimental investigation of the

orthokinetic coalescence efficiency of droplets in simple shear flow, J. Colloid

Interf. Sci. 240, 340–348 (2001).

63. A. Nandi, D. V. Khakhar, and A. Mehra, Coalescence in surfactant-stabilized

emulsions subjected to shear flow, Langmuir 17, 2647–2655 (2001).

64. S. A. Vanapalli and J. N. Coupland, Emulsions under shear—the formation

and properties of partially coalesced lipid structures, Food Hydrocolloids 15,

507–512 (2001).

65. E. P. Schokker and D. G. Dalgleish, The shear-induced destabilization of oil-

in-water emulsions using caseinate as emulsifier, Colloids Surfaces A 145, 61–69

(1988).

66. W. Xu, A. Nikolov, D. T. Wasan, A. Gonsalves, and R. P. Borwankar, Fat

particle structure and stability of food emulsions, J. Food Sci. 63, 183–188

(1998).

67. L. Hamberg, P. Walkenstrom, M. Stading, and A. M. Hermansson,

Aggregation, viscosity measurements and direct observation of protein

coated latex particles under shear, Food Hydrocolloids 15, 139–151 (2001).


10Recent Developments in DoubleEmulsions for Food Applications

Nissim Garti and Axel Benichou

The Hebrew University of Jerusalem, Jerusalem, Israel

I. OPENING REMARKS

Multiple emulsions are emulsion within emulsions. The major types are ofwater-in-oil-in-water (w/o/w) and oil-in-water-in-oil (o/w/o) double emul-sions. These multicompartment liquid dispersions, at least in theory, have asignificant potential in many food applications because the internal dropletscan serve as an entrapping reservoir for active ingredients that can bereleased by a controlled transport mechanism. In practice, double emulsionsconsist of large and polydispersed droplets which are thermodynamicallyunstable with a strong tendency for coalescence, flocculation, and creaming.

Efforts have been made to improve the two major drawbacks of thesesystems related to the thermodynamic emulsion’s instability and the uncon-trolled release of active matter. Almost any possible blends of low-molecu-lar-weight emulsifiers, oils, cosolvents, and coemulsifiers have been tested.The nonviscous fluid double emulsions were always unstable and the releaseof the active matter from the inner phase to the outer continuous phaseremained difficult to control. Only semisolid double-emulsion, gelled orthickened systems have a long shelf life with prolonged stability.Biopolymers, synthetic graft and comb copolymers, and polymerizableemulsifiers that impart steric or mechanical stabilization exhibited improvedstability and better controlled release. Macromolecular surfactants, natu-rally occurring or synthetic polymeric amphiphiles, will increase the viscos-ity of each of the phases, will complex with the oil or the emulsifiers, and will


be able to form systems that will behave much like microcapsules, micro-spheres, and/or mesophasic liquid crystals.

The chapter will stress the most recent findings that can enhance thestability of the double emulsions and/or will reduce droplets sizes for poten-tial food applications. The achievements include (1) choice of food-gradeemulsifiers to enhance emulsion stability at both inner and outer interfaces,(2) droplet size reduction by forming microemulsions (or liposomes) as thevehicles for the active matter in the internal phase, (3) use of differentpreparation techniques to enhance the monodispersibility of the droplets,and (4) use of various additives (carriers, complexing agents, naturalpolymeric emulsifiers) to control and modify the reverse micellar transportphenomena.

Double emulsions were recently used as intermediate structures in thepreparation of microcapsules capable of protecting entrapped addenda andassisting in controlling delivery. The literature is ‘‘flooded’’ with tens of newexamples every year, demonstrating release patterns and control of activeingredients using double emulsions. However, practically no double-emul-sion-based products exist in the marketplace.

The chapter will critically review the relevant literature and will bringsome new emerging improvements involving the stability and the controlissues. Mechanistic considerations will be discussed and alternative ways todeal with the double emulsions concerns related to food applications will beevaluated.

II. INTRODUCTION

Double emulsions are complex liquid dispersion systems known also as‘‘emulsions of emulsions,’’ in which the droplets of one dispersed liquidare further dispersed in another liquid. The inner dispersed globule/dropletin the double emulsion is separated (compartmentalized) from the outerliquid phase by a layer of another phase (1–8).

Several types of double emulsion have been documented. Some con-sist of a single, internal compartment, whereas others have many internaldroplets and are known as ‘‘multiple-compartment emulsions.’’ The mostcommon double emulsions are w/o/w, but in some specific applications,o/w/o emulsions can also be prepared. The term ‘‘multiple emulsion’’ wascoined historically because microscopically it appeared that a number(multiple) of phases were dispersed one into the others. In most cases, itwas proven that, in practice, most systems are composed of double (orduplex) emulsions. Thus, more suitable and more accurate term for suchsystems should be ‘‘emulsified emulsions.’’ A schematic presentation of


the two types of double-emulsion droplets (w/o/w and o/w/o) is shownin Fig. 1.

Potential applications for double emulsions are well documented andmany of these applications have been patented (9–15). In most cases, doubleemulsions are aimed for a slow and sustained release of active matter froman internal reservoir into the continuous phase (mostly water). In someapplications, the double emulsions can serve also as an internal reservoirto entrap matter from the outer diluted continuous phase into the innerconfined space. These applications are aimed at removing toxic matter. Inother applications, double emulsions are reservoirs for improved dissolutionor solubilization of insoluble materials. The materials will dissolve in part inthe inner phase, in part at the internal interface, and occasionally at theexternal interface. Applications related to the protection of sensitive andactive molecules from the external phase (antioxidation) have been recentlymentioned (16–19). Many more applications are expected to emerge in thenear future. Special attention must be paid to the most promising newapplication of double emulsions as intermediate systems in the preparationof solid or semisolid microcapsules (20–22).

III. PREPARATION ROUTES: THE EMULSIFIERS

Double emulsions consist of two different interfaces that require two sets ofdifferent types of emulsifier. In o/w/o double emulsions, the first set ofemulsifiers, for the internal interface, must be hydrophilic and thesecond set of emulsifiers, for the external interface, must be hydrophobic.For w/o/w double emulsions, the order of the emulsifiers is the opposite;

Figure 1 Schematic presentation of the two types of double-emulsion droplet. On

the left is a typical w/o/w multiple droplet and on the right is a typical o/w/o multiple

droplet.


the inner emulsifiers are hydrophobic and the outer ones are hydrophilic. Inmany cases, a blend of two or more emulsifiers in each set is recommendedfor better stabilization results. This review will discuss primarily w/o/wdouble emulsions, as most of the food applications require such emulsions.Some o/w/o emulsions applications are also given.

In early reports on the formation of double emulsions, only one set ofemulsifiers and an inversion process were used. Such preparations were donein one step, but the stability was questionable in most cases. It was difficultto control the distribution of the emulsifiers within the two interfaces. Therewas fast migration of the emulsifiers between the phases that destabilized theemulsions. In most recent emulsion formulations, the emulsions are pre-pared in two steps. At first, a high-shear homogenization was applied onthe water which was added to the solution of the oil and the hydrophobicemulsifiers, to obtain stable w/o emulsion. In the second step, the w/oemulsion is gently added with stirring (not homogenization) to the waterand hydrophilic emulsifier solution (Fig. 2). The droplet size distribution ofa typical classical double emulsion ranges from 10 to 50 mm.

Some more sophisticated preparation methods have been reported inthe literature, of which two are interesting and worth being mentioned. The‘‘lamellar-phase dispersion process’’ was reported by Vigie (23) (Fig. 3). Theprocedure is derived from the process employed to obtain liposomelikevesicles with nonionic emulsifiers. This process can be used only when theconstituents form a lamellar phase by mixing with water in definite propor-tions. This procedure offers some advantage because it requires only asimple emulsification step. The mesophase formed by an ideal ratio of lipo-philic emulsifiers in water is thermodynamically stable and can be obtainedrapidly and easily. The method’s main limitation is derived from the factthat most emulsifiers do not form a lamellar phase. When the lamellarmesophase exists, the hydrophile-lipophile balance (HLB) of the blend ofemulsifiers is often too high, which is disadvantageous for the stability of amultiple emulsion. In addition, the quantity of oil incorporated into thelamellar phase is always low, rarely higher than 10wt%. Another drawbackof this process is the weak control of the rate of encapsulation of the activesubstances.

Grossiord et al. (24) discussed an additional method termed the ‘‘oilyisotropic dispersion process’’ by them. We prefer the more accurate term‘‘emulsified microemulsions.’’ The idea is to disperse an oil phase withinwater by a surfactant and to form an L2 phase (or water-in-oil microemul-sion). This phase is further emulsified with water to form a double emulsion(Fig. 4). The problem is that there is no evidence that the formation of themicrostructures by this method leads to multiple emulsions indeed.Moreover, there is no good evidence that the internal phase remains, after


the second emulsification process, a L2 phase of a submicronal droplet insize. It seems that in some cases, the process is a well-characterized two-stepemulsification that leads to relatively large double-emulsion droplets. Thesame concept of ‘‘emulsified microemulsion’’ was earlier reported by Pilmanet al. (25) and also patented (10). This process is worth further investigationand should be more carefully evaluated. If one can prove that the internalcompartmentalization is of a stable microemulsion, it might bring a break-through to this field because the sizes of the external droplets could bereduced to values below 1 mm. Such formulations will allow the formationof injectable double emulsions.

Gaonkar (10) claimed to have developed a completely new method inthe preparation of double emulsions. This unique new type of preparationdoes not require any homogenization step and does not require the use of a

STEP 1

STEP 2

aqueous phase

w/o emulsion

oil + lipophilicsurfactant

strong

mixing

mixing

w/o emulsion

w/o/wdouble emulsion

water + hydrophilicsurfactants

Figure 2 Schematic of a two-step process in the formation of a double emulsion.


lipophilic emulsifier. In the method, a mixture of oil, water, a second alkylcontaining a polar, protic solvent, such as methanol, ethanol, propanol,glycerol, propylene glycol, dodecanol, and their blend, and a hydrophilicemulsifier is prepared to form a o/w microemulsion. The o/w microemulsionis then diluted with sufficient water to cause destabilization of the micro-emulsion and to provide a w/o/w multiple emulsion.

Higashi et al. (26) described yet another new method of producing w/o/w multiple emulsions by a ‘‘membrane emulsification technique.’’ Thismethod permits the formation of monodispersed liquid microdroplets con-taining aqueous microdroplets to form a w/o/w system. In this method, theaqueous internal phase is mixed with an oil phase containing lipophilicemulsifier. The mixture is sonicated to form a w/o emulsion, which ispoured into the upper chamber of a special apparatus (Fig. 5). The externalaqueous phase containing the hydrophilic emulsifier is continuously injectedinto the lower chamber to create a continuous flow. Nitrogen gas is fed into

Figure 3 Preparation of water-in-oil-in-water multiple emulsion by lamellar-phase

dispersion. (From Ref. 23.)


the upper chamber, initiating a permeation of w/o through the controlled-pore glass membrane into the emulsifying chamber and forming a w/o/wmultiple emulsion. The emulsion is progressively removed from the appara-tus. This process was claimed to be used currently on an industrial scale.

It should be noted that low-molecular-weight emulsifiers migrate fromthe w/o interface to the oil phases and alter the required HLB of each of thephases. Most of the studies, in the years 1970–1985, searched for a propermonomeric emulsifier blend or combination (hydrophilic and hydrophobic)to be used at the two interfaces and the proper ratios between the two.Matsumoto and colleagues (1,27–33) established a minimum ‘‘magic’’weight ratio of 10 of the internal hydrophobic to the external hydrophilicemulsifiers (Fig. 6). Garti and colleagues (34–39) proved that the freeexchange between the internal and the external emulsifiers required a calcu-lation of an ‘‘effective HLB value’’ of emulsifiers to optimize the stabiliza-tion of the emulsion. Complete parameterization work was done on almost

Figure 4 Preparation of water-in-oil-in-water multiple emulsion by ‘‘oily isotropic

dispersion’’ (emulsified microemulsion). (From Ref. 24.)


Figure 5 Preparation of water-in-oil-in-water multiple emulsion by a ‘‘membrane

emulsification technique.’’ (From Ref. 26.)

Figure 6 Yield of formation of double emulsions of w/o/w and o/w/o as a function

of the weight/weight ratio of the internal hydrophobic emulsifier, Span 80, and the

external hydrophilic emulsifier, Tween-80. (From Ref. 1.)


every possible variation in the ingredients and compositions (1–8). In mostcases, the internal emulsifiers were used in great excess to the externalemulsifiers. The nature of the emulsifiers also dictated the number of com-partments and the internal volume that the inner phase occupies.

Many of the more recent studies explore various more sophisticatedemulsifiers such as sphingomyelins (40), modified or purified phospholipids(41), cholates, and so forth. The principles for selecting the properemulsifiers are similar to those known for classical emulsions. Some of theemulsions might have better stability than others, but the general trendremains unchanged. It should be stressed also that one must adjust theemulsifiers to the final application and must substitute one emulsifier bythe other, depending on the total composition of the system.

It must be also recognized that ‘‘empty’’ double emulsions will behavedifferently from those containing active matter (electrolytes, biologicallyactive materials, proteins, sugars, drugs, etc.) due to osmotic pressure gra-dients (caused by the additives) between the outer and the inner phases. Inaddition, many of the active ingredients have some hydrophobicity andsurface properties. Such molecules (peptides, drugs, pesticides) will migratefrom the inner bulk and will adsorb onto the interface, changing the delicateemulsifiers’ HLB. The emulsifiers around the water or the oil droplets willnot cover the droplets fully and the stability will be reduced.

IV. THE OIL PHASE

In food applications, only a limited number of different ‘‘oil phases’’ (water-immiscible liquids) have been suggested and tried throughout years ofresearch. In most applications, the oil phase is based on vegetable oranimal unsaturated triglycerides such as soya oil, cotton, canola, sunflower,and others. It was also suggested to replace the long-chain triglycerides(LCT), which are ‘‘oxygen and hydrolysis sensitive’’ by medium-chain tri-glyceride (MCT), which are fully saturated and thus oxidation resistant.Moreover, the MCT is easier to emulsify and requires less shear.

In cosmetic applications, the freedom to use different oils is greater.Long-chain fatty acids, fatty alcohols, and simple esters such as isopropyl-myristate (IPM), jojoba oil, and essential oils are only few examples. Inaddition, various waxes, sterols, and paraffin oils have been tried.

Several scientists tried to correlate the nature of the oil phase and itsvolume fraction to the stability of the double emulsions. No unusual orsurprising findings were observed. Double-emulsion interfaces behave verymuch like simple emulsions except for the severe limitations on sizes of thedroplets and the internal distribution of the emulsifiers.


V. STABILITY CONSIDERATIONS

Double emulsions made of low-molecular-weight emulsifiers (the so-calledmonomeric emulsifiers) are mostly unstable thermodynamically mainlybecause in the second stage of the emulsification, sever homogenization orshear are not recommended and, as a result, large droplets are obtained.During years of research, attempts have been made to find proper and moresuitable combinations of emulsifiers to reduce droplets sizes and to improvethe emulsion stability. Aggregation, flocculation, and coalescence (occurringin the inner phase and between the double-emulsion droplets) are majorfactors affecting the instability of the emulsions, resulting in rupture ofdroplets and separation of the phases.

Double emulsions are usually not empty. Water-soluble active mate-rials are entrapped during the emulsification in the inner aqueous phase. Itis well documented that because of the difference in osmotic pressurethrough a diffusion-controlled mechanism, the active matter tends to dif-fuse and migrate from the internal phase to the external interface, mostlythrough a mechanism known as ‘‘reverse micellar transport’’ (Fig. 7a). Thedilemma that researchers were faced with was how to control the diffusionof water molecules as well as the emulsifier molecules and mostly theactive matter from the internal phase to the outer phase. It seemedalmost impossible to retain the active material within the water phaseupon prolonged storage. Attempts to increase the HLB of the externalemulsifier or to increase its concentration in order to improve the stability

Figure 7 Schematic of two possible transport mechanisms: (a) reverse micellar and

(b) lamellar thinning transport of a marker from the inner aqueous phase to the

continuous aqueous phase.


of the emulsion worsens the situation and ended in a faster release of thedrug or electrolytes.

Much work was devoted to establish the effects of osmotic pressuredifferences between the internal and the external phases on the stability ofthe emulsions and on the release rates of the markers from the internalphase, and the engulfment of the internal droplets by the flow ofwater from the outer continuous phase to the inner droplets. Single- andmultiple-compartment emulsions were prepared and evaluated in view ofthe enormous potential that these low-viscosity liquid systems have in theslow delivery of water-soluble drugs.

Additional instability mechanisms and release pathways have beendemonstrated and discussed in detail by various authors. These mechanismsinclude ‘‘transport through thinned lamella’’ (Fig. 7b), transport of adductsor complexes that are formed in the oil phase, and other variations of thesemechanisms. It seems however, that the main instability and release mechan-isms are parallel or simultaneously occurring phenomena of ‘‘reversemicellar transport’’ and coalescence.

All of the above mechanisms have been well established, but it seemsthat the stability and the release patterns of these complex double-emulsionsystems depend on various parameters that simultaneously interplay andthat a simplified or unique mechanism cannot explain all of the in-parallelpathways that take place in the double emulsions.

In a recent study, Ficheux et al. (42) identified two types of thermo-dynamic instability that are responsible for the evolution of double emul-sions. Both mechanisms are in good agreement with the Bancroft rule(Fig. 8) but stress different aspects of the previously mentioned mechanisms.The mechanisms elucidated by the authors result in a different behavior ofthe entrapped matter in the double emulsion. The first is a ‘‘coalescence ofthe small inner droplets with the outer droplets interface’’ that is due to therupture of the thin nonaqueous film that forms between the external con-tinuous phase and the inner small water droplets. This instability irreversi-bly transforms a double droplet into a simple direct emulsion. Such amechanism is suitable for delivery of entrapped water-soluble substances.The second mechanism was termed ‘‘coalescence between the small innerdroplets within the oil globule.’’ The first type of instability leads to acomplete delivery of the small inner droplets toward the external phase,whereas the second one does not. The second mechanism leads to anincrease of the average diameter of the internal droplets and a decrease intheir number. The authors worked both with anionic (SDS, sodium dodecylsulfate) and cationic (TTAB, tetradecyltrimethylamonium bromide). It wasdemonstrated that the kinetics associated with the release of the small innerdroplets due to the former instability is clearly related to the hydrophilic


surfactant concentration in the external phase (Figs. 9a and 9b). Dependingon the value of this concentration, double emulsions may be destabilizedwith a timescale ranging from several months to a few minutes.

Rosano et al. (43) explored the influence of ‘‘ripening and interfacialinteractions’’ on the stability of the w/o/w double emulsions. The oil-inso-luble solute was shown to stabilize both the first w/o emulsion (of theinner water droplets) and the external o/w interface. The authors provideda theoretical model and experimental results (video-microscope observa-tions). It was shown that the presence of an electrolyte in the inner waterphase is necessary for the stability of a multiple emulsion. The stability isachieved from the osmotic pressure equalization derived from the differ-ences (excess) in the Laplace pressure. This effect stabilizes the inner w/oemulsion. It is possible also to determine the correct salt concentrationnecessary to balance the osmotic pressure between the two water phases.In a set of experiments (Tables 1 and 2), they show a total of 15 formula-tions in which both the oil phase and the W2 phase are constant, and theonly parameter varied is the NaCl concentration in the W1 phase. The firstsix formulations were prepared with betaine, and the others were madewith SDS. In the case of betaine, without any salt in the inner phase anunstable multiple emulsions is observed and both Ostwald ripening and

PrimaryDispersion

Internaldr opgrowth

Primarydispersion

fluidloss

coalescence

X

Expulsionof

internaldr ops

(or marker: X)

Internal dropgrowth

Primary dispersion

fluid loss

coalescence

Expulsion of

internal drops

(or marker: X)X

Figure 8 Schematic representation of the possible pathways for breakdown in

multiple emulsions.


release of the W1 phase into the W2 phase occur. As the concentration ofthe salt is increased in the W1 phase, the systems do not separate, thestructure is still multiple, and a large increase in viscosity is observed(9500 cP for 0.07wt% to 27600 cP for 0.4wt% NaCl). The authors con-clude that one must consider that three possible factors influence thestability: the Laplace and osmotic pressure effects between the two aqu-eous phases, the interaction between the low- and high-HLB emulsifiers atthe outer o/w interface, and the polymeric thickener–hydrophilic emulsifierinteractions in the outer phase (W2).

Figure 9 (a) Plot, at 20�C, of the lifetime �, of internal droplets entrapped in the oil

globules as a function of the external phase surfactant concentration Ce. The double

emulsions are composed of 90% external phase and 10% double droplets. There is

10% water within the large double globules, 2% Span 80 was used within the oil, and

SDS was used in the external water phase. (b) Influence of the internal surfactant

concentration Ci on the �¼ f(Ce) curve, at 20�C. System: Span 80/SDS as in (a).

(After Ref. 42.)


The variations between the different suggested mechanisms are notdramatic. Some authors tend to stress certain mechanistic aspects and toneglect others, whereas other authors stress, in very specific formulations,the more relevant pathways. It seems that most suggested mechanisms arebasically very similar.

VI. RELEASE CONSIDERATIONS

Several attempts have been made to explain the transport phenomena ofaddenda and water from the inner to the outer phase of w/o/w double-emulsion droplets. It was demonstrated that for un-ionized lipid-solublematerial dissolved in the oil phase, the release obeys first-order kineticsand is diffusion controlled with excellent accordance to Fick’s law.However, ionized and un-ionized water-soluble materials (dissolved in theaqueous inner phase) can also be transported. Matsumoto et al. (27,31) havesuggested two possible mechanisms for the permeation through the oilphase; the first being via the ‘‘reverse micellar transport’’ (Fig. 7a) andthe second by ‘‘diffusion across a very thin lamellae’’ of the surfactantphase formed in areas where the oil layer is very thin (Fig. 7b).

Our studies (37–39) on the release of electrolytes in the presence ofmonomeric emulsifiers, have indicated that the transport takes place even ifthe droplets are very stable to coalescence and even when the osmotic pres-sure of the two phases has been equilibrated. The Higuchi approach and

Table 1 The Role of the Concentration of Salt in the Inner W1 Phase for

Formation Made with Cocamidopropylbetaine

Composition (wt%)

Formulation

number 1 2 3 4 5 6

W1H2O 27.5 27.43 27.35 27.2 27.1 27.2

NaCl 0 0.07 0.15 0.3 0.4 0.3

Light paraffin 21.5 21.5 21.5 21.5 21.5 21.5

Abil EM 90 1 1 1 1 1 1

W2H2O 48.9 48.9 48.9 48.9 48.9 48.9

Betaine 0.3 0.3 0.3 0.3 0.3 0.3

Carbapol 1342 0.1 0.1 0.1 0.1 0.1 0.05

Separation (%) 50 (20)a 0 (80) 0 (80) 0 (80) 0 (80) 0 (110)

Structure Simple Multiple Multiple Multiple Multiple Multiple

aNumber of days.


Table 2 The Role of the Concentration of Salt in the Inner W1 Water Phase for Formulations Made with SDS

Composition (wt%)

Formulation number 7 8 9 10 11 12 13 14 15

W1H2O 27.43 27.4 27.36 27.2 27.1 25.95 26.9 26.6 27.2

NaCl 0.07 0.1 0.14 0.3 0.4 0.45 0.6 0.9 0.3

Light paraffin 21.5 21.5 21.5 21.5 21.5 21.5 21.5 21.5 21.5

Abil EM 90 1 1 1 1 1 1 1 1 1

W2H2O 48.9 48.9 48.9 48.9 48.9 50 48.9 48.9 48.9

Betaine 1 1 1 1 1 1 1 1 1

Carbapol 1342 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2

Separation (%) 4 (40)a 3 (40) 17 (40) 38 (40) 9 (40) 16 (40) 52 (90) 40 (40) 0 (20)

Structure Multiple Multiple Multiple Multiple Multiple Multiple Multiple Multiple Multiple

aNumber of days.


model for the release of matter from polymeric matrices was adopted (44).A model was worked out and a modified release equation was offered andtested. The release factor B was plotted against the time, t, and reciprocalinitial concentrations of the solute (1/C0).

B ¼3

2½1� ð1� FÞ2=3�� F ¼

3Det

r20C0

where De is the effective diffusion coefficient, r0 is the radius of the outerphase droplets, F is the release fraction, and C0 is the initial soluteconcentration.

The straight lines (with accordance with the above expression) and theexcellent correlation coefficient obtained in these experiments indicate theexistence of a ‘‘diffusion-controlled release mechanism.’’ It has been demon-strated that the water can flow into the inner w/o droplets if an osmoticpressure gradient is provided in the presence of hydrophobic emulsifierspresent in the oil phase (32,33). It was also demonstrated that surfactantmolecules, water molecules, and water-soluble addenda are transported atenhanced rates, even without the osmotic pressure gradient, in the presenceof increasing concentrations of a lipophilic emulsifier added to the oil phase(37–39,44). Therefore, it has been concluded that the dominant diffusion-controlled mechanism is facilitated by the presence of reverse micellesformed at the oil phase. The mechanism was termed ‘‘diffusion-controlledrelease mechanism of reverse micellar transport.’’ Most of the releasemechanistic studies in the literature involve double emulsions stabilizedwith synthetic polymeric amphiphiles and therefore, is of less interest forfood applications. Nevertheless, these models of release patterns allow oneto better understand the transport phenomena of entrapped addenda fromthe inner phase to the outer phase.

VII. STABILIZATION BY MACROMOLECULAR AMPHIPHILES

Macromolecules adsorb onto interfaces and facilitate more (or better) cover-age than monomeric emulsifiers. The amphiphilic macromolecules form, inmost cases, thick, flexible films which are strongly anchored into the dis-persed and the dispersion phases. The adsorbed polymers are known toenhance steric stabilization mechanisms and were proved to be good emul-sifiers in food colloids, primarily, in some food-grade oil-in-water emul-sions. The use of macromolecular amphiphiles and stabilizers, such asproteins and polysaccharides, was long ago adopted by scientistsexploring the stability of double emulsions. Gelatin (45,46), whey proteins


(47), bovine serum albumin (BSA) (48–51), human serum albumin (HSA),caseins, and other proteins were mentioned and evaluated. The proteinswere used usually in combination with other monomeric emulsifiers (49).A significant improvement in the stability of the emulsions was shown whenthese macromolecules were encapsulated onto the external interface. Inmost cases, the macromolecule was used in low concentrations (maximum0.2wt%) and in combination with a large excess of nonionic monomericemulsifiers. Furthermore, from the release curves, it seems that the markertransport is more controlled. Dickinson et al. (50–53) concluded thatproteins or other macromolecular stabilizers are unlikely to completelyreplace lipophilic monomeric emulsifiers in double emulsions. However,proteins in combination with stabilizers do have the capacity to confersome enhanced degree of stability on a multiple-emulsion system and,therefore, the lipophilic emulsifier concentration is substantial reduced.

The authors of this review (49) have used BSA along with monomericemulsifiers, both in the inner and the outer interfaces (in low concentrationsof up to 0.2wt%) and found significant improvement both in the stabilityand in the release of markers as compared to the use of the protein in theexternal phase only (Fig. 10). It was postulated that although BSA has nostability effect at the inner phase, it has a strong effect on the release of themarkers (mechanical film barrier). On the other hand, BSA together with

Figure 10 Percent release of NaCl with time from a double emulsion prepared

with 10wt% Span 80 and various BSA concentrations in the inner phase and 5wt%

Span 80þTween80 (1 : 9) in the outer aqueous phase. (From Ref. 49.)


small amounts of monomeric emulsifiers (or hydrocolloids) serves as goodsteric stabilizers and improves stability and shelf life, and slows down therelease of the markers. Therefore, BSA plays a double role in the emulsions:film former and barrier to the release of small molecules at the internalinterface, and steric stabilizer at the external interface. The release mechan-isms involving reverse micellar transport were also established (Fig. 11).

In more recent studies (54,55) the biopolymer chitosan was used as anemulsifier in food double emulsions. Chitosan has surface activity and seemsto stabilize w/o/w emulsions. Chitosan reacts with anionic emulsifiers suchas SDS at certain ratios to form a water-insoluble complex that has strongemulsification capabilities. Chitosan solution was used to form double emul-sions of o/w/o as intermediates from which, by a simple procedure of strip-ing the water, the authors formed interesting porous spherical particles ofchitosan (55).

Cyclodextrins (a, b, and g) were shown to be potential stabilizers for o/w/o emulsions (56). The advantages of the cyclodextrins are their ability tocomplex with certain oil components at the oil–water interface resulting inno need for additional surfactant. It appears that the stabilizer efficacydepends on the nature of the oil and the type of the cyclodextrin(a> b> g). The presence of any active matter in the inner phase (such asbenzophenone) destabilized the emulsion. Only the a-cyclodextrin yieldedstable emulsions. The reason is an interfacial interaction between the com-ponents present at the interface, which changes the HLB and causes adestabilization effect. This elegant idea of interfacial complexation betweenthe ‘‘oil components’’ and the surfactant cannot be a universal solution. Theidea suffers from very severe intrinsic disadvantages once a differentadditive is included in the emulsion. For every additive at any concentra-tion, an adjustment must be made and the given cyclodextrin or thecomplexing agent might be totally unfit.

Figure 11 Schematic of a possible organization and stabilization mechanism of

BSA and monomeric emulsifiers (Span 80) at the two interfaces of a double

emulsion. (From Ref. 49.)


VIII. STABILIZATION BY HYBRIDS OF BIOPOLYMERS

A. Protein–Polysaccharide Interactions in Aqueous Medium

Proteins, polysaccharides, and their blends, as examples of natural biopo-lymers, are surface-active materials. Under specific conditions (such asprotein-to-polysaccharide ratio, pH, ionic strength, temperature, mixingprocessing), it has been stated that proteins and polysaccharides formhybrids (complexes) with enhanced functional properties in comparison tothe proteins and polysaccharides alone. In the case of uncomplexed blendsof biopolymers, competitive adsorption onto hydrophobic surfaces isgenerally reported. Conversely, electrostatic complexation between oppo-sitely charged proteins and polysaccharides allows better anchoring of thenewly formed macromolecular amphiphile onto oil–water interfaces (57).

Recently, the term ‘‘water-in-water emulsion’’ was employed todescribe such dispersions when ‘‘protein-rich aqueous aggregates’’ are dis-persed into a ‘‘polysaccharide-rich aqueous medium’’ (58). A mixture ofgelatin and gum arabic solutions was reported (upon reducing the pH ofthe mixture) to separate in two phases (the upper phase very low in polymercontent and the lower phase containing a highly concentrated precipitate ofthe two polymers). During the last century, many scientists developedmodels and theories to describe polymer interactions and phase separation(incompatibility) which occur when mixing two polymer solutions. Theinteractions between the polymers (protein–protein, protein–polysaccharide,or polysaccharide–polysaccharide) in solution and with the solvent (water inthis case) will govern the solubility and cosolubility of the biopolymers, theviscoelastic properties of the final mixture, and even their behavior atdifferent interfaces (solid–liquid or liquid–liquid).

Natural amphiphilic macromolecules or biopolymers are mainly basedon proteins and polysaccharides. Attraction and repulsion are the two majortypes of interaction that occur between proteins and polysaccharides insolution and can result in complex formation or immiscibility of the twobiopolymers (thermodynamic incompatibility). Owing to polyelectrolyteinteractions in solution, these interactions and their consequences on themixture will be strongly influenced by pH, ionic strength, conformation,charge density, and the concentration of the biopolymers.

In the aqueous solution, complex coacervation takes place betweentwo oppositely charged polymers owing to electrostatic attraction. Forinstance, complexation between proteins and anionic polysaccharidesoccurs below the protein isoelectric point and at low ionic strengths (59).Factors that influence compatibility and complex formation are protein/polysaccharide ratio, pH, ionic strength (60–70), and the nature of thepolymers (molecular weight, net charge, ternary structure, and flexibility


of chains) (70–75). Pretreatment of the polymer’s solutions also enhancesthe complex formation. High-pressure (dynamic or hydrostatic) treatmentas well as temperature have been reported to affect the stabilization of thenewly formed complexes (64,76–79).

Phase diagrams could be built to describe the interactions for eachbiopolymer pair (Fig. 12, right). Two phase-separation phenomena can beobserved, depending on the affinity between the biopolymers and thesolvent. When the Flory–Huggins parameter, �23, which describes the pro-tein–polysaccharide interactions, is positive, indicating a net repulsionbetween the two biopolymers, segregative phase separation (thermodynamicincompatibility) is generally observed. Solvent–biopolymer(s) interactionsare favored to the detriment of biopolymer–biopolymer and solvent–solventinteractions, and the system demixes in two phases, each of which containsboth of the two biopolymers. When the protein–protein or polysaccharide–polysaccharide interactions are sufficiently high in comparison to thepolymer–solvent interactions (i.e., �23<0), congregative phase separation(or coacervation) is observed, leading to an upper solvent-rich phase and alower biopolymer-rich phase forming the so-called ‘‘coacervate.’’ This ispossible at a pH higher than the isoelectric point (IEP) of the protein,when both the protein and the polysaccharide carry the same charge (inthe case of a negatively charged polysaccharide). By reducing the pH ofthe mixture below to the IEP of the protein, the interactions between thetwo oppositely charged biopolymers are favored (�23<0) and complexationtakes place. The newly formed protein–polysaccharide complex could besoluble or can lead to an aggregative phase separation (region II, Fig. 12).The Flory–Huggins theory presents some limits to the protein–polysacchar-ide interactions in aqueous medium-modified models have been establishedto describe these systems in the case of globular proteins and colloidalprotein particles (80,81).

Complexation between two charged biopolymers is usually a reversibleprocess depending on pH and ionic strength. At high ionic strength (0.2–0.3)or at pH values above the protein IEP, the complex will dissociate.However, electrostatic interactions between anionic polysaccharides andpositively charged proteins can occur at a pH above the IEP of the protein.

B. Protein–Polysaccharide Interactions in Emulsion

Protein–polysaccharide interactions play a significant role in the structureand stability of many processed foods. The control of these macromolecularinteractions is a key factor in the development of novel food processes andproducts as well as in the formulation of fabricated products. Much waswritten on the possible interactions between proteins and hydrocolloids in


Figure 12 Phase diagram of water–whey protein isolate (WPI)–xanthan gum

mixtures (top) which describes the different types of interaction between the protein

and the polysaccharide in an aqueous medium (bottom). Binodal curves (solid line)

delimit the boundaries between one phase region (below) and two separated phases

region (above). Region II corresponds to protein–hydrocolloid hybrids formation.


model and real systems but very little advantage was provided by thesestudies in designing new amphiphilic molecules based on these hybrids.The chemically bound structures as well as the hydrogen-bound associatesor even the ion–dipole or dipole–dipole associates form a well-balancedamphiphilic structures which can be utilized in fabricated emulsions andcan replace the synthetic low-molecular-weight emulsifiers. These associateswill not require any additional legislation (i.e., FDA) and will have texturaland stability advantages. If one can add some nutritional or health benefitsto it (and the hydrocolloids have shown such effects), the fabricated emul-sions, food products, or food processes will become more attractive to thefood producers.

Dickinson and colleagues claimed (82,83) that the best way to adsorbhydrocolloids onto interfaces is to link them to proteins. The proteins aresurface-active materials consisting of flexible hydrophobic and hydrophilicmoieties, preferentially adsorbing onto the interfaces and replacing thehydrocolloids from the surface. A hydrocolloid that is thermodynamicallyincompatible with an adsorbed protein (58,60) can be distributed at theinterface only if it interacts somehow with the protein. Its distribution onthe interface will depend on the nature of these interactions. Strong chemicalbonds will cause a different surface distribution rather than weak associa-tions between the two.

It was also recognized that polysaccharides could interact at the inter-face with other polymers (nonproteins, other polysaccharides) as well aswith groups residing at the interface (protein–water, oil–water, or air–water) with the formation of an aqueous structured material with usefulviscoelastic mechanical properties under conditions of low shear stress.Understanding the detailed chemistry of these polymer–polymer interac-tions and relating them to the observed rheology is a significant aspect ofthe study of hydrocolloids.

Complexation between proteins and polysaccharides at the emulsiondroplet surface can improve steric stabilization. Droplet size can be smallerif the polysaccharide is present during homogenization, and, therefore, therate of creaming may be reduced so long as there is no bridging flocculation.Covalent protein–polysaccharide complexation can also provide effectiveemulsion stabilization. The great improvement in stability arising fromthe presence of the polysaccharide during emulsification is attributable tothe formation of a thicker, stronger steric-stabilizing layer around thedroplets.

Tolstoguzov et al. (84) compared the surface properties of BSA andBSA–dextran complexes and the stability of n-decane-in-water emulsionsprepared with them was evaluated by measuring the volume of the coexist-ing phases obtained after phase separation, under centrifugation (50min at


23,000 g). Complete phase separation occurred for emulsions prepared with0.2wt% BSA. When BSA–dextran complexes were used (pH 6.0, 0.3wt%biopolymer concentration), only 40% of the decane separated after centri-fugation. Because the emulsification properties of the complexes werestrongly dependent on pH and ionic strength, the authors attributed theemulsification power of BSA–dextran complexes to the electrostaticnature of the interactions. Even though both biopolymers carry a net nega-tive charge at pH 7 (BSA–dextran sulfate, 1 : 3 by weight), a soluble ioniccomplex can be formed via local electrostatic interaction between the highlycharged anionic polymer and positively charged patches on the globularprotein. Surface shear viscosity measurements give independent evidenceof an interfacial complex between BSA and dextran sulfate.

A similar synergistic effect of protein–polysaccharide complexes instabilizing emulsions has been reported with corn oil–water emulsions pre-pared with casein–acidic polysaccharide complexes (85). For a pectin-to-total biopolymers ratio >0.2, an emulsion stability index of 100% wasobtained corresponds to zero creaming of the emulsion after centrifugationfor 30min at 3000 g. Adsorbed complexes at the oil–water interfaceincreased the interfacial viscosity, forming a gellike structure around theoil droplets. Moreover, these emulsions exhibited relatively high thermalstability at 95�C because of the gellike microstructures at the interface.Emulsions prepared with soy proteins–sodium alginate complexes at pH7.0 and for 10 : 1 and 4 : 1 protein-to-polysaccharide ratios (86) exhibitedimproved thermal stability in comparison to soy protein-stabilized emul-sion. The emulsion activity index (EAI) values were 0.801 and 0.953 for acontrol emulsion (standard emulsion stabilized with soy protein) and for anemulsion prepared with soy proteins–alginate complexes (protein-to-poly-saccharide ratio of 4 : 1), respectively. The improvement of the emulsionstability and activity indices was attributed to the unfolding of the soyprotein at the interface that enhances protein–polysaccharide interactionsby exposing cationic groups and hydrophobic groups. The newly exposedhydrophobic groups could alter the complex surface properties by providingadditional sites for binding lipids, and that allows better anchoring of theamphiphilic macromolecules into the oil phase, resulting in an enhancedemulsion activity and stability indices.

Nevertheless, under specific conditions of pH, ionic strength, andpolysaccharide concentration, protein–polysaccharide complexes in emul-sions could contribute, in some cases, to a decrease in the emulsion stability.Dickinson and Pawlowsky (87) envisaged the flocculation, creaming, andrheology properties of n-tetradecane-in-water emulsions stabilized withBSA–i carrageenan complexes. The authors envisaged first the surface prop-erties of soluble complexes by determining the evolution in time of the


surface tension [g(t)] at different pH values (Fig. 13). BSA always providedlower g values at any pH. The higher values of g obtained with BSA–i-carrageenan complexes have been attributed to a diminution of thenumber of free BSA molecules at the interface owing to complex coacerva-tion, to a smaller diffusion coefficient as a result of complex sizes, and toincreased viscosity of the solution. The electrostatic nature of BSA–i-carra-geenan complexes was clearly shown and could explain the dependence ofthe emulsion stability on pH and ionic strength. A relatively poor stability ofthe emulsion was described for polysaccharide concentrations above0.01wt% and was attributed to bridging flocculation of the oil droplets.At high carrageenan content (>0.1wt%), stability against creaming of theemulsions was increased by the formation of a polysaccharide cross-linkednetwork droplets with gellike rheological properties.

Covalently linked complexes have also shown emulsification capability(88). Because such complexes did not present much interest for food appli-cations, we describe them succinctly. Oleic acid–water emulsions preparedwith covalent linked b-lactoglobulin–carboxymethyl dextran conjugateswere unaffected by heat treatment up to 80�C, demonstrating the heatstability of the complexes. Dickinson et al. (89) compared the influence of

Figure 13 Time-dependent surface tension (g) of biopolymer solution (5mM

imidazol, 25�C): e and f pH¼ 8.0; i and m pH¼ 6.5; œ and g pH¼ 6.0. Open

and closed symbols refer respectively to 10�3 wt% BSA and 10�3 wt% BSAþ 4.10�3

wt% i-carrageenan. (From Ref. 87).


ionic and covalent protein–polysaccharide interactions on 10wt% n-hexane-in-water emulsions stabilized with BSA, nonionic dextran, and anionic dex-tran sulfate at neutral pH (Fig. 14). Both BSA–dextran conjugates (curve A)and BSA–dextran sulfate conjugates (curve C), yielded stable emulsionswith droplet diameters that remained below 1.5 mm after 3 weeks standingat 25�C. Similar emulsion stability was obtained using a BSA–dextran sul-fate mixture (curve D). On the other hand, emulsions prepared with a BSA–dextran mixture were unstable. Conversely conjugating or mixing BSA withamylopectin was not sufficient to stabilize emulsions (curves E and F).

Several scientists envisaged the preemulsification process influence onthe emulsifying properties of protein–polysaccharides complexes (90–93).Hydrostatic pressures and preheating of the biopolymers suspensions havebeen often reported as a useful tool to modify and, to some extent, toenhance the emulsification power of these mixtures.

Heating was reported to increase the surface hydrophobicity of glob-ular proteins such as ovalbumin, but a decrease in surface hydrophobicitywas also demonstrated for BSA and b-lactoglobulin (88,91). Conversely, thehydrophobicity of 11S Vicia faba globulin was significantly increased byheating the native protein solution to 80�C for 2 min (93). When polysac-charide (i.e., i-carrageenan) was added to untreated 11S Vicia faba globulinor heated at 80�C for 2min, the surface hydrophobicity of the protein

Figure 14 Average droplet diameter d43 as a function of time t of emulsions stored

at 25�C and stabilized with: (i), BSA-dextran conjugate (curve A) (m), BSA

dextran mixture (curve B) (r), BSA–dextran sulfate conjugate (curve C); (n), BSA

dextran sulfate mixture (curve D); (e), BSA-amylopectin conjugate (curve E); (f),

BSA amylopectin mixture (curve F). (From Ref. 89).


remained unchanged. The temperature treatment of the protein solutioninduced extensive precipitation, whereas the mixed solution remainedclear. It was concluded that i-carrageenan inhibited the aggregation of theprotein molecules by protecting protein-reactive sites.

The influence of temperature treatment on the emulsifying efficiencyof 11S Vicia faba globulin pure or mixed with i-carrageenan or k-carragee-nan in 20 wt% n-tetradecane-in-water emulsions was also studied (93). Heattreatment was reported to lower the emulsifying power of the protein. Tothe contrary, polysaccharide addition (protein to polysaccharide ratio of 3.3to 1) leads to emulsions with smaller droplets at all temperatures. Mixedemulsions prepared with i-carrageenan (high content of sulfate groups) havebeen found to be more stable against creaming than those prepared withk-carrageenan (less sulfated). The average droplet diameters of emulsionswere of 5.7 and 8.0 mm, respectively, for mixtures heated for 2min at 80�C.High-pressure treatment affects significantly the surface hydrophobicity ofproteins and protein–polysaccharide blends (90–93). High pressure inducessubunit dissociation (quaternary structure) and, to some extent, unfoldingof the globular subunits which result in the exposure of new hydrophobicgroups within the protein molecules (93). Polysaccharide addition reducedthe surface hydrophobicity of the pressurized protein by blocking the hydro-phobic site onto the protein surface. In spite of reduced hydrophobicity,mixtures of pressurized k-carrageenan–protein and i-carrageenan–proteinlead to the formation of stable emulsion with respect to creaming. Thesame trend as for temperature treatment was observed. The droplet sizesof emulsions prepared with the i-carrageenan–protein mixture were smallerthan those prepared with the k-carrageenan–protein blend. Conversely,increasing the applied pressure on the protein solution yielded emulsionswith larger droplets. This was attributed to protein aggregation by disulfidebinding between free-cystein residues.

C. Double Emulsion Stabilized withProtein–Polysaccharide Hybrids

Recently, we have envisaged the stabilization of double emulsions with some‘‘molecular-recognition’’ hybrid of whey protein isolate (WPI) and differentcharged and uncharged polysaccharides (7,94). The emulsifying power ofWPI–xanthan gum hybrids was studied. It was found that WPI–xanthanratio strongly influences the formation and stability of the droplets(Fig. 15). At low protein content (0.5–2wt%), the droplet size decreaseswith time due to internal aqueous-phase expulsion. This was attributed tothe weakness of the emulsifier film adsorbed onto the external oil–waterinterface. Microscopic observation revealed that for a WPI–xanthan gum


ratio of 0.5/0.5 (w/w), double-emulsion droplets inverted to simple o/wdroplets (empty of internal globules) after aging for 28 d at room tem-perature. At higher protein-to-polysaccharide ratios (3/0.5 to 6/0.5, w/w)double-emulsion droplets exhibited increased stability against creaming andcoalescence. Double emulsions were prepared with different whey protein–xanthan ratios. Increasing the xanthan content to 1wt% decreases the dro-plets diameter from 11.8 to 3.1 mm (Fig. 16a). It was found that by increasingthe protein-to-polysaccharide ratio (region II of the phase diagram), double-emulsion droplets with reduced droplet size and increased stability areformed.

One can conclude that the binodal boundary line corresponds to thelimit of complexation of the two biopolymers that will give optimum stabil-ization. At biopolymer content higher than the binodal composition, theaddition of one of the two biopolymers will only slightly affect the double–emulsion droplets’ stability and will only add a small depletion stabilizingeffect to the emulsion droplets.

1. Rheology

Rheological measurements were performed on double-emulsion droplets.The phase angle, d, was defined as arctan(G00/G0), where G0 is the storagemodulus, G00 is the loss modulus, and tan(d)¼G00/G0. It was found that at

Figure 15 Double-emulsion droplet stability with time at room temperature

prepared with 0.5wt% xanthan gum and increasing amount of WPI as an external

emulsifier: 0.5wt%, ^; 1wt%,œ; 2wt%,i; 3wt%, �; 4 wt%, þ�; 5wt%,e;

6wt%,þ.


low levels of gum (0.1wt%), the emulsions had high viscosity, with thephase angle (d) close to 90�, indicating self-assembly of the two biopolymersonto the external oil–water interface. Therefore, the emulsions are stabilizedmainly by steric interactions between the ‘‘macromolecular-recognitionhybrids’’ adsorbed onto the oil. To the contrary, at a high level of gum(0.3–1wt%), the emulsions exhibit more elasticity, regarded as a physicalproperty derived from a depletion stabilization mechanism. The protein willpreferably adsorb onto the oil–water external interface and the uncom-plexed gum will migrate to the bulk and contribute to the stabilization bya depletion mechanism (Fig. 16b). The protein-to-gum ratio of 4/0.5, atwhich complexation between the protein and the gum takes place,

Figure 16 (a) WPI–xanthan gum ratio (w/w) influence on the droplet diameter

(mm) of double-emulsion droplets, immediately after preparation. (b) WPI–xanthan

gum ratio (w/w) influence on the phase angle d (degrees) of double-emulsion

droplets, immediately after preparation. The phase angle, d, was defined as

arctan(G0 0/G0), where G0 is the storage modulus, G0 0 is the loss modulus, and

tan(d)¼G0 0/G0.


corresponds to an intermediary viscoelasticity of the system. Microscopicobservations and zeta-potential measurements performed on the samplesafter rheological studies revealed that the emulsion droplets are still multipledroplets and that interactions between the biopolymeric molecules remainedunchanged.

At a gum concentration of 0.5wt%, the protein concentration doesnot affect the rheological behavior of the double emulsion that conserves itselasticity properties at all ratios with phase angle (d) values around 25� at allprotein contents. At the same time, the yield of preparation of doubleemulsions prepared with increasing amounts of protein and at 0.5wt%xanthan gum was 64% for double emulsions stabilized with 6wt% WPIand increased to 96% with the addition of 0.5wt% xanthan gum.

2. Effect of pH on Double-Emulsion Stability

The pH is known to influence strongly the net harge of proteins. Therefore,it was important to qualify and to quantify its effect on the protein–poly-saccharides interactions in solution and at the oil–water interface in doubleemulsions. The zeta potential of both solutions and emulsions was deter-mined (Fig. 17). For the WPI–xanthan gum system, the pH affects theprotein, the gum, and the combination of the two. The blends behave asthe protein alone with very close isoelectric points to pure protein. Once thehybrid is adsorbed onto the oil droplets, the situation is significantly differ-ent. The isoelectric point of the protein-stabilized emulsion decreases, indi-cating unfolding of the protein and screening of its charged residues during

Figure 17 Zeta potential (mV) of solutions (left) and of multiple emulsions (right)

prepared with 4wt% WPI (œ), 0.5 wt% xanthan gum (i), and the blend of the two

at a weight ratio of 4/0.5 (e).


the emulsification process. At a low pH, one can conclude that the protein isdominant at the interface, and at a high pH, the gum is the dominant one atthe interface. In other words, when the pH is reduced to below the isoelectricpoint of the blend, the steric stabilization is dominant, and when the pHincreases, the system is mainly stabilized by a depletion mechanism. The pHalso influences strongly the rheological behavior of double emulsions. At apH below the isoelectric point of the blend [WPI–xanthan gum (4/0.5)], 4.8,the phase angle, d, is higher than 45�, indicating that the emulsion isviscous and confirming steric stabilization mechanism for these systems.Conversely, at a higher pH (5–10), the system is predominantly elastic (dclose to 25� in all the pH range), indicating a depletion stabilization mechan-ism. Microscopy observations of emulsion stabilized with a WPI–xanthangum mixture (4/0.5) at pH 7 reveals a weakness of the film around the oilglobules, indicating the presence of uncomplexed biopolymers at the inter-face. It was concluded that at a pH below the isoelectric point of the blend,the two biopolymers strongly interact and yield fully covered double-emul-sion droplets. When the pH of the external aqueous phase increases, at aconstant WPI–xanthan gum ratio, the electrostatic interactions betweenthe protein and the gum become weaker around the oil droplets and the‘‘biomolecular-recognition hybrid’’ film becomes less rigid.

3. Entrapment in Double Emulsions (Vitamin B1)

Much work was done on the entrapment of addenda in double emulsions,but only a few reports have shown promising results on real systems. Inthis subsection, we will bring a more recent example drawn from theauthors’ work.

The formulations stabilized with the mixture of WPI and xanthan gumwere used for entrapping vitamin B1 in the inner aqueous phase. A usefulelectrochemical method was developed to quantify the amount of vitaminB1 present in the external aqueous phase. Differential pulse polarographybased on the redox properties of the vitamin when it is present in theexternal aqueous phase allowed the determination of its release profilesunder various conditions of pH, protein-to-polysaccharide ratio, and initialconcentration of the entrapped addenda. The release profile, as a function ofthe pH of the external aqueous phase (Fig. 18) reveals that at pH 2, theexternal interface is better sealed against the release of the entrapped vita-min than at pH 4 and 7. It reconfirmed that there is an extra stability ofthese double emulsions derived from both electrostatic and steric mechan-ism stabilization of the external oil–water interface.

At pH 3.5, the increase in the protein-to-polysaccharide ratio alsoreduces the release rate. As was demonstrated earlier, at high hydrocolloid


content, the system is stabilized by both electrostatic and steric mechanismsthat improve the rigidity of the film around the double-emulsion globulesand so delay the vitamin release (Fig. 19).

IX. STABILIZATION BY SOLID PARTICLES

To assist in the stabilization of emulsions, particles (such as fat crystals)must be adsorbed at the emulsion droplet interface, providing a physicalbarrier to coalescence. It is known that in many emulsified foods, solidparticles are necessary for producing stabilization (e.g., ice crystals in icecream, egg yolk particles in mayonnaise, and fat particles in whipping

Figure 18 Effect of the external aqueous phase pH on the release profile of vitamin

B1 from multiple emulsions stabilized with WPI–xanthan gum (4/0.5) as the external

emulsifier (pH 7,f; pH 4,m; pH 2,g).

Figure 19 Release profile of vitamin B1 from multiple emulsions at different

WPI–xanthan ratios as the external emulsifier at pH 3.5 (4/0, œ; 4/0.1, m; 4/0.2,

^; 4/0.5, f).


cream). The key factors that will determine the influence of fat crystals onemulsion stabilization are (1) the wettability of the crystals at the interface,(2) interfacial film rheology, (3) particle microstructure (polymorphism andmorphology), and (4) location of fat crystals [(in the dispersed (o/wemulsion) or continuous phase (w/o emulsion)] (95).

The wetting behavior of particles at the interface is described by con-tact angles, which are related to the surface tension of each of the threeinterfaces by Young’s equation: g (o/w)cos y¼ g (o/s)� g (w/s), where y is thecontact angle measured through the water phase and g (o/w), g (o/s), and g (w/s)

are the surface tensions of the oil–water, oil–solid, and water–solid inter-faces, respectively. cosy g (o/w) is also known as the adhesion tension.Modification of the contact angle (and, therefore, emulsion stability) canby achieved by a modification of the aqueous, oil, or solid phase so as toalter g (o/w), g (o/s), or g (w/s).

In Fig. 20, where water acts as the continuous phase and oil as thedispersed phase, the colloidal particle is partially embedded within the inter-face at an equilibrium distance dictated by the interparticle forces. With acontact angle less than 90� at the solid–water–oil interface across the waterphase, the particle will stabilize o/w emulsions. With contact angles greater

Figure 20 Adsorption and contact angles of fat crystals at the interface of an

oil-in-water emulsion.


than 90�, the particles will stabilize a w/o emulsion. If the particles used arecompletely wetted by either the oil or water phase (and, therefore, thecontact angle is 180�), they become fully dispersed in that phase and anyemulsion-stabilizing effect is negated. A 90� contact angle means that acrystal is equally wetted by the oil and aqueous phases.

Stabilization of margarine and other similar food emulsions isachieved by partial adsorption of solid fat particles (b0-polymorphs) ontothe water–oil interface, bridged by monomeric hydrophobic emulsifiers. Thecomplex stabilization is achieved by ‘‘wetting’’ the oil phase by solid fatparticles and emulsifiers (lecithins and monoglycerides of fatty acids). Theconcept was recently reexamined and reconsidered by Bergenstahl (96,97)and the mechanism was somewhat better elucidated.

The concept of stabilizing emulsions by solid particles (mechanicalstabilization) was partially demonstrated in an old and incomplete study(98), showing that colloidal microcrystalline cellulose (CMCC) particlescan adsorb in a ‘‘solid form’’ onto oil droplets at the interface of water–oil emulsions and thus improve their stability by mechanical action. Oza andFrank (98) tried the mechanical stabilization concept on double emulsions.The report shows some promise in improving stability and in slowing downthe release of drug. This study was, for several years, the only example ofapplying the concept of mechanical stability on double emulsions. In a mostrecent article, Khopade and Jain (99) repeated the use of a similar processand managed to stabilize w/o/w emulsions by using MCC (microcrystallinecolloidal cellulose) particles at both interfaces. The droplets were small andthe yield of the double emulsion was fairly good. The increasing concentra-tion of MCC in either the internal or external phase increased droplet sizes.These systems showed promise in tuberculosis therapy.

Garti et al. (100) carried out several experiments with micronizedparticles of the a- and b0-polymorphs of tristearin fat together with poly-glycerol–polyricinoleate (PGPR) as the internal emulsifiers in double emul-sions. Solid fat particles did not sufficiently stabilize the water-in-oilemulsion, and, similarly, the PGPR (at the concentrations used in the for-mulation) did not provide good stability. However, it was shown that theuse of the blend of the two components composed of a solid submicronal fatparticles of a- and b0-polymorphs (which are more hydrophilic than the b-form and thus wet the oil–water interface better) ‘‘precipitate’’ onto thewater droplets and ‘‘cover’’ them. The fat particles bridge between thewater droplets and sinter them only if a lipophilic surfactant (PGPR) wascoadsorbed onto the water–oil interface (the w/o emulsion) (Fig. 21). Theauthors’ interpretation of the results is that ‘‘the fat particles adsorb ontothe hydrophobic emulsifier film and both, the solid particles and the emul-sifier, wet the water and spread at the interface.’’ The double emulsions


prepared by this technique were more stable than those prepared bymonomeric emulsifiers.

Organophilic montmorillonite is an interesting clay that gained someinterest in the emulsion technology. Stable o/w/o emulsions with compo-nents consisting of hydrophilic nonionic surfactant (hydrogenated, ethoxy-lated castor oil, HCO-60), organophilic montmorillonite, and commercialnonionic surfactant (DIS-14) were made (101). The montmorillonite wasadded in the second step at the outer w/o/w interface. The droplets sizesdecreased with the increase of the HCO-60 (0.1–3wt%) concentration. Theviscosity of the double emulsion increased as the concentration of themontmorillonite, and DIS-14 increased, indicating that the excess amountof the inner oil phase is adsorbed by the outer oil phase. The results indicatethat the weight fraction of the inner oil phase should not exceed 0.3wt% forstable o/w/o emulsions because the viscosity of the double emulsion is sohigh that the formulation becomes semisolid.

X. STABILIZATION BY INCREASED VISCOSITY

Most food emulsions are highly complex systems both in terms of composi-tion and structure (102). To control the formation, the stability and rheol-ogy of food emulsions require an understanding of the interactions betweenthe various elements present in the system. Recently, progress has beenmade to understand the interaction behavior at the o/w interface between

Figure 21 A Schematic of a colloidal margarine structure demonstrating the role

of emulsifiers and fat crystals in stabilizing the (w/o) emulsion of margarine by

colloidal fat crystals. (From Ref. 100.)


some of the components found in food emulsions, particularly betweendifferent proteins and surfactants, and, to a lesser extent, between proteinsand fat crystals. Because fat crystal interactions at the interface are affectedby protein, and protein interfacial behavior is affected by small-moleculesurfactants, it is important to examine how all of these components interact.As with bulk emulsion properties, the interface can exhibit viscous, elastic,and viscoelastic properties. The ability of an emulsion to resist coalescencewill largely depend on the properties of the interface. A highly viscous andrigid interfacial film will retard the rate of film drainage and resist rupture,thereby promoting stability (103). Hence, by controlling interface rheology,one can control the drainage of thin liquid films trapped between coalescingdroplets, which may also affect the displacement of crystals away from theinterface during droplet coalescence. Kiosseoglou (104) discussed the role ofsurface-active lipids at the o/w interface on the viscoelastic parameters ofBSA, sodium caseinate, and egg yolk films during adsorption at the oliveoil–protein solution interface. It was postulated that film viscoelasticity wasthe result of surface-active lipids interacting with folded and unfoldedproteins at the interface.

Restricting the mobility of the active matter in the different compart-ments of the double emulsion will slow down coalescence and creaming, aswell as decrease the transport rates of the drug, or the marker, from thewater phase through the oil membrane. Attempts were made to do thefollowing: (1) increase the viscosity of the internal aqueous phase byadding gums/hydrocolloids to the inner water phase, such a thickenermay affect also the external continuous phase because the entrapment isnot quantitative and the yields of entrapment are limited and are emulsifierdependent, (2) increase the viscosity of the oil phase (fatty acids salts); (3)increase the thickening or gelling of the external water by gums. This processis limited to cosmetic (or similar) applications in which semisolid emulsionsare directly applied (Tables 3 and 4). Some of these examples are topical skincare products, creams, and body lotions (105,106).

Double emulsions that were solidified after preparation may sufferfrom destabilization effects. This phenomenon is scarcely considered butin practice it can occur very often. The solidification occurs because oftemperature changes (temperatures can fluctuate from sub-zero of ca�20�C to ca 40�C) during transport or storage. Clausse et al. (107,108)studied the phenomena in w/o/w emulsions by microcalorimetric [differen-tial scanning calorimetry (DSC)] techniques. It was concluded that, becauseof thermodynamic equilibrium, double emulsions may suffer from watertransfer during the solidification. This phenomenon occurs even if partialsolidification takes place. In addition, a change in the size distributionof emulsion droplets is observed. The mean diameter of the droplets in the


w/o emulsion may shift toward the o/w emulsion and the double emulsioncan invert. Therefore, it is not always obvious that increasing viscosity,gelation, or partial solidification improved the emulsion’s stability.

XI. MICROCAPSULES OR MICROSPHERES IN THEINTERNAL PHASE

Entrapping the active matter in solid or semisolid particles will dramaticallydecrease their release rates. Such double emulsions can be stored, before use,for prolonged periods of time without transporting the active matter to theouter interface. Upon use, the double emulsion will be heated or shearedand the solid internal matrix will be ruptured and the active matter shouldbe released. The major problem in practicing such technology is the

Table 3 Example of w/o/w Multiple Emulsion Stabilized by a Calcium Alginate

Gel Layer at the Internal Water Oil Interface

Oil phase

Hexaglycerol mixed ester 1%

CSL (calcium stearyl-2-lactylate) 0.75%

Soybean oil 23.25%

Internal aqueous phase 27% Solution sodium alginate, low viscosity 25%

Outer aqueous phase Polysorbate 20 0.36%

Water 9.14%

Vinegar 12%

Sucrose 6%

NaCl 2.5%

Aqueous phase þ 1% xanthan gum 20%

Table 4 Composition for Low-Fat Spread Using a w/o/w Formulation

Outer aqueous phase Salt 1%

Gelatin 3%

Maltodextrin DE ¼ 7a 10%

Water 85%

Sodium caseinate 1%

w/o emulsion Water 57.9%

Rapeseed oil 40%

PGPR 2%

Flavor 0.1%

aDE¼ dextrose equivalent.


difficulties arising in dispersing (and in keeping it stable) the micropraticleor nanoparticles in the continuous water phase. Microspheres and nanopar-ticles using solid encapsulation techniques were tested to replace the classicalw/o emulsion. Only a few experiments were carried out showing that releasecan be slowed down, but the stability of these systems is very limited (109–123). Such methods are applicable only for emulsions that can be freshlyprepared prior to their use. It is our hope that more efforts will be made inthis direction.

XII. DOUBLE EMULSIONS AS INTERMEDIATES FORNANOPARTICULATION

Advanced double-emulsion formulations are no longer prepared for thepurpose of simple delivery and release of active matter but have changedapplication directions. Three main new directions can be clearly seen: (1)double emulsions as intermediates for the preparation of solid microspheresor miocrocapsules, (2) o/w/o double emulsions for improved solublizationand chemical protection of water-insoluble active matter, and (3) doubleemulsions for selective adsorption of certain compounds for extractionand purification purposes.

Some major examples along with the classical delivery applications inpharmaceuticals and food applications will be described that will emphasizethe new emerging applications.

A. Controlled Delivery Applications

The intrinsic instability of the double emulsion caused difficulties for for-mulators and many of the investigators have decided to abandon this tech-nology (6,7). However, one should bear in mind that the potentialapplications for double emulsions appear to be enormous, mainly in theareas of food, cosmetics, medicine, and pharmaceutics. Throughout theyears, potential applications have been demonstrated in (1) improved bio-logical availability (parenteral nutrition, anticancer drugs), (2) delivery ofdrugs (sustained release of narcotic antagonistic drugs, prolonged release ofcorticosteroids, slow and targeted release of anticancer drugs), and (3)adsorption of toxic compounds. The technology has much promise also innonpharmaceutical areas of slow and controlled release of materials such asfertilizers and pesticides for agricultural formulations as well as for con-trolled release for cosmetic, industrial, and food applications (6,7). In fact,most of the applications of double emulsions for controlled delivery are inthe field of pharmaceuticals and health.


Double emulsion may offer some advantages for food applicationmainly with relation to their capability to encapsulate (or entrap) in theinternal compartments some water-soluble substances which are thenslowly released. The double emulsion can also be used in the food industrywhere an external water phase is more acceptable in terms of palatabilitythan an oil one (124,125). On this basis, several new products have beenpatented in the form of w/o/w emulsions, such as salted creams (encapsula-tion of salt), aromatic mayonnaise, and so for in (126–129). Further foodapplications are related to the double-emulsion dielectric properties. Forexample, one can prepare a w/o/w system having the same volume fractionof the dispersed phase and the same texture as a simple emulsion, but with alower oil content (due to the existence of the aqueous compartments in thefood globules) (i.e., low-calorie mayonnaise) (124).

B. Double-Emulsions Solvent-Extraction Techniques forPreparation of Microspheres

Drugs, cosmetic ingredients, and foods additives are microencapsulated forvariety of reasons, which include reducing local side effects, controlledrelease, site-specific (drug) delivery, and drug targeting. A tremendousamount of research work is done in a search of suitable methods to achievea good encapsulation of water-soluble active matter. The physical charac-teristics of the microspheres produced largely determine their suitability foruse for different objectives. Microspheres are prepared from both naturaland synthetic polymers.

Among microencapsulation techniques, the double-emulsion solvent-evaporation method is one of the most useful methods for entrapping water-soluble compounds (109–123). Fig. 22 shows schematically the preparationtechnique. Over 100 articles and many patents have been published in thecourse of the last 5 years on the use of this technique. Most of the applica-tions are rather related with pharmaceuticals and drug encapsulation thanwith foods. We have selected only some examples that are of a more sig-nificant value.

The w/o/w emulsions are generally used for encapsulating proteins orpeptides. These highly water-soluble molecules are quantitatively introducedin the internal aqueous phase of the multiple emulsions and result in anincreased encapsulation efficiency microcapsules in comparison to particlesproduced by single emulsion solvent-evaporation method. The particularlocation of the proteins induce a stabilizing effect on the two emulsions,which, in turn, contributes to a successful stabilization of the doubleemulsion and loading.


The double-emulsion solvent-evaporation technique is commonly usedto prepare biodegradable hydrophobic microspheres containing hydrophilicpharmaceuticals, proteins, and polypeptides for sustained release applica-tions (111,115–123). In most cases, the microspheres are in the range size of10–100 mm. However, recently, Blanco-Prieto et al. (116) managed to reducethe microcapsules sizes to less than 5 mm.

Couvreur et al. (115) reviewed the preparation and characterization ofmany of the different types of the solvent-evaporation microsphere andmostly discussed small poly(lactic-co-glycolic acid) microspheres (meansize lower than 10 mm) containing small peptides (Fig. 23). Three mainevaporation strategies have been utilized in order to increase the encapsula-tion capacity an interrupted process, a: continuous process, and the rotaryevaporation procedure.

Much work was devoted in recent years to preparing microparticles ofnarrow size distribution with different biodegradable polymers. The sizes ofcommon microcapsules are 40–50 mm (115–116,121–123).

Liquid–liquid emulsification is a critical step in the double-emulsionmicroencapsulation process (w/o/w or o/w/o). It was found that the size ofthese droplets decreases with increasing homogenization intensity and dura-tion. The emulsion droplet size depends, as expected, on viscosity, totalvolume size, and volume ratio of the continuous phase to the dispersed

Figure 22 Preparation of microspheres by the multiple-emulsion solvent-evapora-

tion technique.


phase: the rotor/stator design was also investigated. All of these physicalparameters influence the structure of the microspheres obtained by thistechnique.

An interesting example is the incorporation of a protein-based drug inmicrospheres made of a hydrophobic polymer via double liquid–liquidemulsification (w/o/w) or by dispersing a powdered protein in a polymersolution followed by liquid–liquid emulsification (s/o/w) (117). BSA wasused as the model protein and poly(methyl methacrylate) (PMMA) wasused as the model polymer. The droplets sizes of the w/o emulsion dropletswere controlled using rotor/stator homogenization. The size of the micro-spheres thus prepared was found to increase with the increasing size of theprotein powder in the s/o/w system but increase with decreasing size of the

Figure 23 Release kinetic of pBC 264 in phosphate-buffered saline pH 7.4: (top) in

vivo and (bottom) in brain tissue from PLGA microspheres prepared with either

ovalbumin (f) or Pluronic F-68 (œ). (From Ref. 115).


liquid emulsion droplets in the w/o/w system. Empirical correlation canaccurately predict the size of the microspheres. Protein loading in the micro-spheres decreased with respect to increase in w/o emulsion droplet size or inprotein powder size.

In a most recent article (122), it was demonstrated that a milk modelprotein, b-lactoglobulin (BLG), was encapsulated into microspheres pre-pared by the solvent-evaporation technique. The effect of the pH of theouter aqueous phase on the protein encapsulation and release as well ason the microsphere morphology has been investigated. It was demonstratedthat as the amount of BLG increases, the stability of the inner emulsiondecreased and the entrapment was less efficient. Therefore, adjusting thecombined pH effect and the stability of the inner emulsion may lead tobetter entrapment. As to the release, it was demonstrated that the ‘‘bursteffect’’, attributed to a morphology change in the microcapsules character-ized by the presence of pores, or channels, able to accelerate the release ofthe BLG, is the most significant release factor. These pores were attributedto the presence of a large amount of BLG on the surface which aggregatesduring microsphere formation at pH 5.2. It was concluded that it is bene-ficial to lower the solubility of the protein in the outer phase in order toimprove the encapsulation efficiency, although this benefit is provided by astrong adsorption of the protein on microsphere surface. Some more inter-esting articles on protein entrapments for various oral and other intakes canbe found in the literature (117–120). Chitosan porous spherical particleswere prepared from o/w/o double emulsions stabilized with a chitosanaqueous solution. The particulation was obtained by a simple evaporationtechnique (55).

XIII. O/W/O DOUBLE EMULSIONS

Oil-in-water double emulsions were considered to have less potentialapplications and therefore were less extensively studied. However, in morerecent years, several new applications have been mentioned for o/w/odouble emulsions which sound interesting and worth being mentioned.

Modulated release of triterpenic compounds from a o/w/o multipleemulsion formulated with dimethicones studied with infrared spectrophoto-metric and differential calorimetric approaches is one of these examples(130). The authors explored the advantages in the release of triterpeniccompounds from o/w/o emulsions. They found two principal advantages:(1) the use of low-molecular-weight silicones decreased the oily touch of thefinal preparation and (2) due to the large range of viscosity, this excipientinfluenced the skin distribution of the active matter after the topical


application. The effects of different triterpenic compounds incorporatedwithin multiple emulsions were studied, through in vitro penetration results.The residual film on the skin was also evaluated. Correlations wereestablished between the silicone structure and the distribution of drugs indifferent skin levels or between the silicone structure and the percutaneouspenetration. The incorporation of silicones within o/w/o multiple emulsionsseems to be an efficient means of modulating the penetration and thedistribution of drugs in the skin.

In another study, the stability of retinol (vitamin A alcohol) was com-pared in three different emulsions: oil-in-water (o/w), water-in-oil (w/o), andoil-in-water-in-oil (o/w/o) (16). The stability in the o/w/o emulsion was thehighest among the three types of emulsion. The remaining percentages, at50�C after 4 weeks, were 56.9, 45.7, and 32.3, in the o/w/o, w/o, and o/wemulsions, respectively. However, it was also reported that with the increas-ing peroxide value of o/w and w/o emulsifiers, the remaining percentage ofvitamin A palmitate and retinol in the emulsions increased significantly,indicating that peroxides in the formulas accelerate the decomposition ofvitamin A. An organophilic clay mineral tan oil gelling agent and a w/oemulsifier also affected the stability of retinol. The stability of retinol in theo/w/o emulsion increased with increasing inner oil-phase ratio, whereas ino/w. it was unaffected by the oil fraction. The encapsulation percentage ofretinol in the o/w/o emulsion, the ratio of retinol in the inner oil phase to thetotal amount in the emulsion, increased with increasing oil fraction. Theremaining percentage of retinol in the o/w/o emulsion was in excellentagreement with encapsulation percentage, suggesting that retinol in theinner oil phase is more stable than that in the outer oil phase. Addition ofantioxidants (tert-butylhydroxytoluene, sodium ascorbate, and EDTA) tothe o/w/o emulsion improved the stability of retinol up to 77.1% at 50�Cafter 4 weeks. The authors concluded that the o/w/o emulsion is a usefulformula for stabilizing vitamin A.

Orange oil-in-water emulsions were encapsulated in another oil phaseto form a double emulsion with orange oil inside its inner compartment (19).Although the yield was only 44.5%, it is a promising area for furtherresearch in the future in order to prevent air oxidation of the oil. Spray-drying of the double emulsion can provide a secondary coating to secure amaximum protection of orange oil and affords a free-flowing flavor powder.Spray-drying of the orange oil double emulsion had a very low destructioneffect on its structure, as revealed by light microscopy. This method mayhave a potential application in different food or pharmaceutical productswhere maximum protection is required. A secondary coating was applied tothe flavor oil already encapsulated in a double emulsion. The spray-dryingtechnique includes spraying a flavor emulsion into a stream of hot air.


The water phase is then evaporated rapidly, leaving the flavor materiallocked in the carrier.

XIV. MULTIPLE-EMULSION RHEOLOGY

The understanding of the rheological behavior of double emulsions is quiteimportant in the formulation, handling, mixing, processing, storage, andpipeline transformation of such systems. Furthermore, rheological studiescan provide useful information on the stability and internal microstructureof the double emulsions. Some attention was given to this subject in recentyears and the results are significant because they help to clarify certainaspects on the stability and release properties of the double emulsions(131). Few publications deal with w/o/w double emulsions and only onedeals with the o/w/o emulsion. Most of the rheology work is old and donot contribute a new understanding. Therefore, we will bring only recentrelevant work.

An interesting characterization of the mechanical properties of the oilmembrane in w/o/w emulsions was done by an aspiration technique (132). Itwas adopted from techniques related to the evaluation of global or celldeformability. The deformability of an individual globule during a totalor partial flow into a cylindrical glass tube calibrated under well-controlledconditions of aspiration was determined. An analysis of the behavior of thedouble emulsion by a migration of the lipophilic surfactant to the interfacebetween the oily and the external aqueous phases was done. It was shownthat the elastic shear modulus and the interfacial tension of the oilymembrane increased with the lipophilic surfactant concentration. Thisstudy also provides an explanation of the mechanism related to theswelling–breakdown process from physical and mechanical considerations.

Grossiord et al. (131–133) applied linear shear flow on the w/o/wdouble emulsions that contained active matter, and from the rheologicalpatterns, they learned about the bursting effect of the droplets with therelease of entrapped substances and the composition of the system. Theauthors (133) described a set of two types of experiment: oscillatory dynamictests and a steady-state analysis. They measured the stress and strain of theemulsions by applying sinusoidal shear. These parameters (shear or complexmodulus G*; the lag phase between stress and strain, d; the storage modulusG0 and the loss modulus G00) provide a quantitative characterization of thebalance between the viscous and elastic properties of the multiple emulsions.At lag phase, d¼ 0�, and when ¼ 90�, the system is viscoelastic. The shearsweep and the temperature sweep characterize the multiple emulsion at rest.Fig. 24 describes a transition between an elastic behavior and a viscous


behavior, which occurs at critical stress values. The change in this par-ameters indicate a pronounced structural breakdown. The authors consid-ered the influence of the formulation parameters on the swelling and releasekinetics using the rheological properties. Parameters such as the oil nature,the width of the oil membrane, and the lipophilic and the hydrophilic natureof the surfactant have been evaluated. The two main parameters identifiedto affect the swelling/breakup kinetics were the difference in concentrationin water soluble molecules between the internal and the external aqueousphase (Fig. 25), and the lipophilic surfactant concentration. It was alsoobserved that the maximum viscosity values increase with the surfactantratio (under hypo-osmotic conditions, Fig. 26). The same trend wasobserved when the release of water-soluble materials, b(t), was followed(Fig. 27). The authors concluded that a progressive migration of theexcess lipophilic surfactant in the oil phase toward the primary or thesecondary interface occurs.

Stroeve and Varanasi (134) also examined the breakup of the multiple-mulsion globules in a simple shear flow and concluded from the criticalWeber number (We)cr (Figs. 28 and 29) that the multiple emulsions exhibitbehavior that is similar to that of simple emulsions. From Fig. 28, onecan see, at least qualitatively, the evolution of (We)cr as a function of p

Figure 24 Changes in G*, G0 0 and d for increasing stress at fixed frequency in

double emulsions. The double-emulsion composition of the water oil emulsion is

24% oil, various lipophilic surfactant concentrations and 0.7% MgSO4. The o/w/o

emulsion contain 80% water in oil, 2% hydrophilic surfactant, and demineralized

water. (From Ref. 133.)


Figure 26 Change in viscosity versus time under hypo-osmotic conditions for

different lipophilic surfactant concentrations. (From Ref. 133.)

Figure 25 Change in viscosity versus time for different concentration gradients in

the inner phase. (From Ref. 133.)


Figure 27 Change in released fraction b(t) versus time under hypo-osmotic

conditions for different lipophilic surfactant concentrations. (From Ref. 133.)

Figure 28 Changes in the critical Weber number (We)cr for simple- and multiple-

emulsion disruptions as a function of the viscosity ratio dispersed to the continuous

phase. (From Ref. 134.)


(the viscosity ratio between the drop and the continuous phase) for simpleand multiple emulsions and that there are some differences between the twoemulsions. The double emulsion has more heterogeneous characteristics.From Fig. 29, it is possible to obtain the minimum shear rate value whichis able to produce the breakup of the oil globules and the release of water-soluble encapsulated molecules.

The studies showed also that the mechanisms taking place during thebreakup were complex and did not always lead to a total release of theentrapped electrolyte. Some phenomena such as a partial leakage ofthe internal aqueous compartment or the expulsion of the aqueousmicroglobules covered by a residual lipophilic film were able to restrictthe release.

De Cindio and Cacace (125) prepared food double emulsions andstudied their rheological behavior by steady shear and oscillatory measure-ments. They have concluded that the w/o/w appeared to have rheologicalproperties examined to those of a simple o/w emulsion having the samefraction of dispersed phase but lower oil content (Fig. 30). It was alsodemonstrated that the plot of both storage moduli G0 and G00 versusoscillation frequency, W, are similar in all eight prepared emulsions, andthe loss tangent is about 1, and both elastic and viscous contributions to theviscoelastic behavior of double emulsions are of similar magnitude. Thesimilarity of texture between simple and double emulsions is very encoura-ging, leading to some interesting conclusions and new perspectives. Theinfluence of the mixture of emulsifiers on the double-emulsion stability

Figure 29 Changes in the critical Weber number (We)cr for multiple-emulsion

disruption as a function of the viscosity ratio dispersed to continuous phase, for

different primary emulsion volume fractions. (From Ref. 134.)


was studied by an oscillatory ring-surface rheometer from which theinterfacial elasticity at the oil–water interface can be evaluated.

Pal (135) studied the rheology of o/w/o double emulsions. The simpleo/w emulsions were found to be Newtonian up to a dispersed-phase con-centration of 45% by volume and to be non-Newtonian above this volumefraction. All of the double o/w/o emulsions are highly non-Newtonian. Thedegree of shear thinning increases with the increase in primary o/w emulsionconcentration (Fig. 31). The oscillatory measurements indicate that themultiple emulsions are predominantly viscous in that the loss modulusfalls above the storage modulus over the entire frequency range investigated(Fig. 32). Upon aging, the storage and loss moduli of the double emulsionsshow a significant increase. However, the increase in viscosity with aging isonly marginal.

The rheological behavior of the w/o/w emulsion studied under a conicplate viscometer has shown a negative thixotropic flow pattern, mostlyunder a low shear rate. Upon raising the shear rate or the shear time, anincrease of the shear stress was observed, which induced phase inversion to aw/o of a semisolid-type emulsion. The hydrodynamic parameters (dissipatedenergy, kinetic energy, and impulse applied to the emulsion by the rotatingcone) causing the phase inversion were determined and a mechanism for

Figure 30 Apparent viscosity versus shear rate for both w/o/w and o/w systems at

volume fraction of disperse phase of 0.3. (From Ref. 125).


Figure 32 Storage and loss moduli for multiple o/w/o emulsions as functions of

frequency. (From Ref. 135.)

Figure 31 Apparent viscosity as a function of shear stress for multiple o/w/o

emulsions. (From Ref. 135.)


such an inversion was suggested. Fig. 33 shows a proposed mechanism forphase inversion under the shear rate (136). Induced shear, causing phaseinversion, could serve as a possible technique to release the drug.

XV. CONCLUDING REMARKS

Double emulsions are known for over three decades and were extensivelystudied in the last 15 years. The internal phase is an excellent reservoir foractive matter that needs protection and can be released in a controlled rate.However, the sizes of the droplets and the thermodynamic instability were asignificant drawback of this technology.

The use of conventional low-molecular-weight emulsifiers did notsolve these problems. However, much progress was made with the introduc-tion of amphiphilic macromolecules as emulsifiers. This multianchoring offlexible macromolecules can improve the steric stabilization by forming, athick, multilayered coating on the droplets. A variety of hybrids, complexes,adducts, and others between the amphiphiles and coemulsifiers or cosol-vents has been suggested. These molecules improved significantly the stab-ility and slowed the release rates. Physical methods of separation, filtration,and extraction also had a positive effect on the release patterns of any drugor active matter. Progress was made also in identifying and characterizingthe parameters and mechanisms involved in coalescence, aggregation, andrupture of the double-emulsion droplets. A good control of the rheological

Figure 33 Proposed mechanism of phase inversion under shear. (From Ref. 136.)


parameters was achieved by a better understanding of their effect on thestatic and shear-induced stability.

It seems that the double-emulsion technology can now be applied invarious application areas, mainly in food, cosmetics, and pharmaceuticals(for nonintravenous applications).

The mains goals remain to obtain submicronal double-emulsiondroplets with long-term stability (possible by emulsified microemulsions)and to trigger and control the release at will.

Compatible blends of biopolymers (hydrocolloids and proteins) areexcellent future amphiphilic candidates which, under certain combinations,will serve both as ‘‘release controllers’’ and ‘‘stability enhancers’’ for thefuture preparations of double emulsions.

ACKNOWLEDGMENTS

The authors are grateful to Dr. Abraham Aserin for his valuable help andcritical reading of the present manuscript.

This work is dedicated to the late Frida and Karl Kissman andsupported by the ‘‘Kissman Fund for Applied Chemistry,’’ Jerusalem,Israel.

REFERENCES

1. S. Matsumuto, and W. W. Kang, Formation and applications of multiple

emulsions, J. Dispersion Sci. Technol. 10, 455–482 (1989).

2. N. Garti, Double emulsions-scope, limitations and new achievements, Colloids

Surf. A: Physicochemical and Engineering Aspects 123, 233–246 (1997).

3. N. Garti, and A. Aserin, Double emulsions stabilized by macromolecular

surfactants, in Surfactants in Solution, (A. K. Chattopadhyay and K. L. Mittal,

eds.), Surfactant Science Series Vol. 64, Marcel Dekker, New York, 1996,

pp. 297–332.

4. N. Garti, Delivery of microparticulated liquid systems in food, in Handbook of

Non-Medical Applications of Liposomes, (Y. Barenholtz and D. D. Lasic, eds.),

CRC, New York, 1996, Vol. 3, pp. 143–199.

5. N. Garti, Influence of the formulation on the characteristics and stability of

multiple emulsions, in Multiple Emulsions: Structure, Properties and Applications

(M. Seiller and J. L. Grossiord, eds.), Editions de sante, Paris, 1998, pp. 81–116.

6. N. Garti, Progress in stabilization and transport phenomena of double

emulsions in food applications, Food Sci. Technol. 30, 222–235 (1997).

7. N. Garti, and C. Bisperink, Double emulsions: progress and applications, Curr.

Opin. Colloid Interf. Sci. 3, 657–667 (1998).


8. N. Garti and A. Benichou, Double emulsions for controlled-release applica-

tions-Progress and trends, in Encyclopedic Handbook of Emulsion Technology

(J. Sjoblom, ed.), Marcel Dekker, New York, 2001, pp. 377–407.

9. L. Thill-Francis, Stable double emulsions containing finely divided particles,

U.S. patent 93/00007, 1993.

10. A. G. Gaonkar, Method for preparing a multiple emulsion, U.S. patent

5322704, 1994.

11. A. G. Gaonkar, Stable multiple emulsion comprising interfacial gelatinous

layer, flavor-encapsulating multiple emulsions and low/no-fat food products

comprising the same, U.S. patent 5332595, 1994.

12. Y. Takahashi, T. Yoshida, and T. Takashi, Process for producing a W/O/W

type multiple emulsion for medicines, cosmetics, etc., U.S. patent 4985173, 1994.

13. P. Strauel, and G. Friour, Process for preparing photographic emulsions

having a low fog level, U.S. patent 94/420106, 1994.

14. C. A. Herb, Rinse-off water-in-oil-in-water compositions. European patent

0717978A2, 1996.

15. G. W. Bams, and W. H. Van Megen, Edible water-in-oil-in-water emulsion,

U.S. patent 4650690, 1987.

16. K. Yoshida, T. Sekine, F. Matsuzaki, T. Yanaki, and M. Yamaguchi, Stability

of vitamin A in oil-in-water-in-oil-type multiple emulsions, J. Am. Oil Chem.

Soc. 76(2), 195–200 (1999).

17. S. Y. Kim, and Y. M. Lee, Lipid nanospheres containing vitamin A or vitamin

E: Evaluation of their stabilities and in vitro skin permeability, J. Ind. Eng.

Chem. 5(4), 306–313 (1999).

18. M. Gallarate, M. E. Carlotti, M. Trotta, and S. Bovo, On the stability of

ascorbic acid in emulsified systems for topical and cosmetic use, Int. J.

Pharmaceut. 188, 233–241 (1999).

19. A. Edris, and B. Bergnsthal, Encapsulation of orange oil in a spray dried

double emulsion, Nahrung Food 45(2), 133–137 (2001).

20. B. B. C. Youan, J. Gillard, and B. Rollmann, Protein-loaded poly(epsilon-

caprolactone) microparticles III. Entrapment of superoxide dismutase by the

(water-in-oil)-in-water solvent evaporation method, STP Pharma Sci. 9(2),

175–181 (1999).

21. J. L. Grossiord, M. Seiller, I. Pezron, and D. Clausse, Multiple emulsions:

Vesicular systems with industrial applications, Oleagineux Corps Gras Lipides

3(3), 158–162 (1996).

22. T. Uchida, K. Yoshida, and S. Goto, Preparation and characterization of

polylactic acid microspheres containing water-soluble dyes using a novel w/o/w

emulsion solvent evaporation method, J. Microencapsulation 13(2), 219–228

(1996).

23. L. Vigie, Emulsions multiples h/l/h: nouveaux procedes de fabrication et de

caracterisation, Graduate thesis, University Paris XI, 1992.

24. J. L. Grossiord, M. Seiller, and A. Silva-Cunha, Obtaining multiple emulsions,

in Multiple Emulsions: Structure, Properties and Applications (G. Seiller and

J. L. Grossiord, eds.), Editions de sante, Paris, 1998, pp. 57–80.


25. E. Pilman, K. Larsson, and E. Torenberg, Inverse micellar phases in ternary

systems of polar lipids/fats/water and protein emulsification of such phases to

w/o/w–microemulsion–emulsions, J. Dispersion Sci. Technol. 1, 267–281 (1980).

26. S. Higashi, M. Shimizu, T. Nakashima, K. Iwata, F. Uchiyama, S. Tateno,

S. Tamura, and T. Setoguchi, Arterial-injection chemotherapy for hepato-

cellular carcinoma using monodispersed poppy-seed oil microdroplets contain-

ing fine aqueous vesicles of epirubicin, Cancer 75, 1245–1254 (1995).

27. S. Matsumoto, Y. Kida, and D. Yonezawa, An attempt at preparing water-in-

oil-in-water multiple-phase emulsions, J. Colloid Interf. Sci. 57, 353–361 (1976).

28. S. Matsumoto, M. Koda, and S. Murata, Preparation of lipid vesicles on the

basis of a technique for providing w/o/w emulsions, J. Colloid Interf. Sci. 62,

149–157 (1977).

29. S. Matsumoto, Y. Ueda, Y. Kita, and D. Yonezawa, Preparation of water-in-

oil and oil-in-water multiple emulsions in a eatable form, Agric. Biol. Chem. 42,

739–743 (1978).

30. S. Matsumoto, Development of W/O/W-type dispersion during phase inversion

of concentrated W/O emulsions, J. Colloid Interf. Sci. 94, 362–368 (1983).

31. S. Matsumoto, T. Inoue, M. Koda, and T. Ota, An attempt to estimate stability

of the oil layer in water/oil/water emulsions by means of viscometry, J. Colloid

Interf. Sci. 77, 564–565, (1980).

32. S. Matsumoto, Y. Koh, and A. Michiura, Preparation of water/olive oil/water

emulsions in an edible form on the basis of phase inversion technique,

J. Dispersion Sci. Technol. 6, 507–521 (1985).

33. S. Matsumoto, W/O/W-type multiple emulsions with a view to possible food

applications, J. Texture Studies 17, 141–159 (1986).

34. M. Frenkel, R. Shwartz, and N. Garti, Stability, inversion, apparent and

weighted HLB, J. Colloid Interf. Sci. 94, 174–178 (1983).

35. N. Garti, M. Frenkel, and R. Schwartz, Multiple emulsions. Part II: Proposed

technique to overcome unpleasant taste of drugs, J. Dispersion Sci. Technol. 4,

237–252 (1983).

36. S. Magdassi, M. Frenkel, and N. Garti, On the factors affecting the yield of

preparation and stability of multiple emulsions, J. Dispersion Sci. Technol. 5,

49–59 (1984).

37. S. Magdassi, M. Frenkel, N. Garti, and R. Kasan, Multiple emulsions II—

HLB shift caused by emulsifier migration to the external interface, J. Colloid

Interf. Sci. 97, 374–379 (1984).

38. S. Magdassi, M. Frenkel, and N. Garti, Correlation between nature of emul-

sifier and multiple emulsion stability, Drug Dev. Ind. Pharm. 11, 791–798 (1985).

39. S. Magdassi, and N. Garti, A kinetic model for release of electrolytes from

w/o/w multiple emulsions, J. Control. Release, 3, 273–277 (1986).

40. A. J. Khopade, and N. K. Jain, Long-circulating multiple-emulsion system for

improved delivery of an anticancer agent, Drug Delivery 6(2), 107–110 (1999).

41. M. P. Y. Piemi, M. De Luca, J. L. Grossiord, M. Seiller, and J. P. Marty,

Transdermal delivery of glucose through hairless rat skin in vitro: effect of

multiple and simple emulsions, Int. J. Pharm. 171(2), 207–215 (1998).


42. M. F. Ficheux, L. Bonakdar, F. Leal-Calderon, and J. Bibette, Some stability

criteria for double emulsions, Langmuir 14(10), 2702–2706 (1998).

43. H. L. Rosano, F. G. Gandolfo, and J. D. P. Hidrot, Stability of W-1/O/W-2

multiple emulsions—Influence of ripening and interfacial interactions, Colloids

Surfaces A: Physicochemical and Engineering Aspects 138(1), 109–121 (1998).

44. R. G. Stehle, and W. I. Higuchi, In vitro model for transport of solutes in

three phase system. I. Theoretical principles. II. Experimental considerations,

J. Pharmaceut. Sci. 61, 1922–1930 (1972).

45. A. Vaziri, and B. Warburton, Some preparative variables influencing the

properties of w/o/w multiple emulsions, J. Microencapsulation 11(6), 649–656

(1994).

46. W. Zhang, T. Miyakawa, T. Uchida, and S. Goto, Preparation of stable w/o/w

type multiple emulsion containing water-soluble drugs and in-vitro evaluation

of its drug releasing properties, J. Pharm. Soc. Japan. 112(1), 73–80 (1992).

47. M. Cornec, P. J. Wilde, P. A. Gunning, A. R. Mackie, F. A. Husband, M. L.

Parker, and D. C. Clark, Emulsion stability as affected by competitive

adsorption between an oil-soluble emulsifier and milk proteins at the interface,

J. Food Sci. 63(1), 39–43 (1998).

48. E. Fredrokumbaradzi, and A. Simov, Effect of bovine serum albumim (BSA)

on release of sulfacetamide sodium from multiple w/o/w emulsions, Pharmazie

47(5), 388–389 (1992).

49. N. Garti, A. Aserin, and Y. Cohen, Mechanistic considerations on the release

of electrolytes from multiples emulsions stabilized by BSA and nonionic

surfactants, J. Control. Release 29, 41–51 (1994).

50. E. Dickinson, J. Evison, R. K. Owusu, and A. Williams, Protein stabilized

water-in-oil-water emulsions, in Gums and Stabilizers for the Food Industry

(G. O. Phillips, P. A. Williams and D. J. Wedlock, eds.), IRL Oxford Uni-

versity Press, Oxford, 1994, Vol. 7, pp. 91–101.

51. E. Dickinson, J. Evison, and R. K. Owusu, Preparation of fine protein-

stabilized water-in-oil-in-water emulsions, Food Hydrocolloids 5, 481–485

(1991).

52. J. Evison, E. Dickinson, R. K. Owusu Apenten, and A. Williams, Formulation

and properties of protein-stabilized W/O/W multiple emulsion, in Food

Macromolecules and Colloids (E. Dickinson and D. Lorient, eds.), Royal


53. A. Koberstein Hajda, and E. Dickinson, Stability of water-in-oil-in-water

emulsions containing faba bean proteins, Food Hydrocolloids 10(2), 251–254

(1996).

54. P. C. Schulz, M. S. Rodriguez, and L. F. Del Blanco, Emulsification properties

of chitosan, Colloid Polym. Sci. 276(12), 1159–1165 (1998).

55. T. Adachi, J. Ida, and M. Wakita, Preparation of spherical and porous chitosan

particles by suspension evaporation with O/W/O multiple emulsions, Polym. J.

31(4), 319–323 (1999).

56. S. C. Yu, A. Bochot, M. Cheron, M. Seiller, J. L. Grossiord, G. Le Bas, and

D. Duchene, Design and evaluation of an original o/w/o multiple emulsion


containing natural cyclodextrins as the emulsifier, STP Pharm. Sci. 9(3),

273–277 (1999).

57. A. Benichou, A. Aserin, and N. Garti, Protein–polysaccharide interactions for

stabilization of food emulsions. A review, J. Dispersion Sci. Technol. 23(1&3),

93–123 (2002).

58. V. B. Tolstoguzov, Protein–polysaccharide interactions, in Food Proteins

and their Applications (S. Damodaran and A. Paraf, eds.), Marcel Dekker,

New York, 1997, pp. 171–200.

59. V. B. Tolstoguzov, Functional properties of food proteins and role of protein–

polysaccharide interactions, Food Hydrocolloids 4, 429–468 (1991).

60. V. Y. Grinberg, and V. B. Tolstoguzov, Thermodynamic incompatibility of

proteins and polysaccharides in solutions, Food Hydrocolloids 11, 145–158

(1997).

61. C. Sanchez, C. Schmitt, V. G. Babak, and J. Hardy, Rheology of whey protein

isolate–xanthan mixed solutions and gels. Effect of pH and xanthan

concentration, Nahrung 41, 336–343 (1997).

62. P. Walkenstrom, N. Panighetti, E. Windhab, and A. M. Hermansson, Effect of

fluid shear and temperature on whey protein gels, pure or mixed with xanthan,

Food Hydrocolloids 12, 469–479 (1998).

63. S. I. Laneuville, P. Paquin, and S. L. Turgeon, Effect of preparation conditions

on the characteristics of whey protein–xanthan gum complexes, Food

Hydrocolloids 14, 305–314 (2000).

64. C. M. Bryant, and D. J. McClement, Influence of xanthan gum on physical

characteristics of heat denaturated whey protein solutions and gels, Food

Hydrocolloids 14, 383–390 (2000).

65. H. Zaleska, S. G. Ring, and P. Tomasik, Apple pectin complexes with whey

protein isolate, Food Hydrocolloids 14, 377–382 (2000).

66. C. Schmitt, C. Sanchez, F. Thomas, and J. Hardy, Complex coarcevation

between b-lactoglobulin and acacia gum in aqueous medium, Food

Hydrocolloids 13, 483–496 (1999).

67. F. Simonet, C. Garnier, and J. L. Doublier, Partition of proteins in the guar/

dextran two-phase systems, Food Hydrocolloids 14, 591–600 (2000).

68. M. M. Alves, C. Garnier, J. Lefebvre, and M. P. Goncalves, Microstructure

and flow behavior of liquid water–gelatin–locust bean gum systems, Food

Hydrocolloids 15, 117–125 (2001).

69. M. M. Ould Eleya, and S. L. Turgeon, Rheology of k-carrageenan and

b-lactoglobulin mixed gels, Food Hydrocolloids 14, 29–40 (2000).

70. M. M. Ould Eleya, and S. L. Turgeon, The effects of pH on the rheology of

b-lactoglobulin/k-carrageenan mixed gels, Food Hydrocolloids 14, 245–251

(2000).

71. A. Antonov Yu, and M. P. Goncalves, Phase separation in aqueous gelatin-

k-carrageenan systems, Food Hydrocolloids 13, 517–524 (1999).

72. M. Verheul, and S. Roefs, Structure of whey protein gels, studied by

permeability, scanning electron microscopy and rheology, Food Hydrocolloids

12, 17–24 (1998).


73. P. Walkenstrom, and A. M. Hermansson, Effect of shear on pure and mixed

gels of gelatin and particulate whey protein, Food Hydrocolloids 12, 77–87

(1998).

74. P. Walkenstrom, E. Windhalb, and A. M. Hermansson, Shear-induced

structuring of particulate whey protein gels, Food Hydrocolloids 12, 459–468

(1998).

75. F. Delben, and S. Stefancish, Interaction of food polysaccharides with

ovalbumin, Food Hydrocolloids 12, 291–299 (1998).

76. C. M. Bryant, and D. J. McClements, Optimizing preparation conditions for

heat-denatured whey protein solutions to be used as cold-gelling ingredients,

J. Food Sci. 65(2), 259–263 (2000).

77. V. B. Galazka, I. G. Summer, and D. A. Ledward, Changes in protein–protein

and protein–polysaccharide interactions induced by high pressure, Food Chem.

57, 393–398 (1996).

78. P. Paquin, Technological properties of high pressure homogenizers: The effect

of fat globules, milk proteins, and polysaccharides, Int. Dairy J. 9, 329–335

(1999).

79. N. Lagoueyte, and P. Paquin, Effect of microfluidization on the functional

properties of xanthan gum, Food Hydrocolloids 12, 365–371 (1998).

80. Y. T. Wu, D. Q. Lin, and A. Q. Zhu, Thermodynamics of aqueous two-phase

systems. The effect of polymer molecular weight on liquid–liquid equilibrium

phase diagrams by the modified NRTL model, Fluid Phase Equilibria 147,

25–43 (1998).

81. J. L. Doublier, C. Garnier, D. Renard, and C. Sanchez, Protein–polysaccharide

interactions, Curr. Opin. Colloid Interf. Sci. 5, 202–214 (2000).

82. E. Dickinson, The role of hydrocolloids in stabilizing particulate dispersions

and emulsions, in Gums and Stabilisers for the Food Industry (G. O. Phillips,

D. J. Wedlock and P. A. Williams, eds.), IRL Press, Oxford, 1988, Vol. 4,

pp. 249–263.

83. V. B. Galazka, E. Dickinson, and D. A. Ledward, Effect of high pressure on

ovalbumin–polysaccharide interactions. High Pressure Res. 19(1–6), 515–520

(2000).

84. N. A. Larichev, A. N. Gurov, and V. B. Tolstoguzov, Protein–polysaccharide

complexes at the interface. I. Characteristics of decane/water emulsions

stabilized by complexes of bovine serum albumin with dextran sulfate,

Colloids Surfaces 6, 27–34 (1983).

85. E. S. Tokaev, A. N. Gurov, I. A. Rogov, and V. B. Tolstogusov, Properties of

oil/water emulsions stabilized by casein-acid polysaccharide mixture, Nahrung

31, 825–834 (1987).

86. M. S. Lippi, and M. V. Taranto, Soy protein–acidic polysaccharide interaction:

Modification of the emulsification properties of soy protein isolate, Food Sci.

Technol. 14(2), 55–59 (1981).

87. E. Dickinson, and K. Pawlowsky, Effect of �-carrageenan on flocculation,

creaming and rheology of a protein-stabilized emulsion, J. Agric. Food Chem.

45, 3799–3806 (1997).


88. K. Nagasawa, K. Takahashi, and M. Hattori, Improved emulsifying

properties of b-lactoglobulin by conjugating with carboxymethyl dextran,

Food Hydrocolloids 10(1), 63–67 (1996).

89. E. Dickinson, and V. B. Galazka, Emulsion stabilization by protein–

polysaccharides complexes, in Gums and Stabilisers for the Food Industry

(G. O. Phillips, D. J. Wedlock and P. A. Williams, eds.), IRL Press, Oxford,

1992, Vol.6, pp. 351–361.

90. V. B. Galazka, E. Dickinson, and D. A. Ledward, Influence of high pressure

on interactions of 11S globulin vicia faba with i-carrageenan in bulk solution

and at interfaces, Food Hydrocolloids 14, 551–560 (2000).

91. E. Dickinson, and J. D. James, Influence of high-pressure treatment on b-lactoglobulin-pectin associations in emulsions and gels, Food Hydrocolloids

14, 365–376 (2000).

92. E. Dickinson, and K. Pawlowsky, Effect of high-pressure treatment of protein

on the rheology of flocculated emulsions containing protein and polysacchar-

ide, J. Agric. Food Chem. 44, 2992–3000 (1996).

93. V. B. Galazka, E. Dickinson, and D. A. Ledward, Emulsifying Behavior of

11S globulin vicia faba in mixtures with sulphated polysaccharides:

Comparison of thermal and high-pressure treatments, Food Hydrocolloids

13, 425–435 (1999).

94. A. Benichou, A. Aserin, and N. Garti, Multiple emulsions stabilized by new

molecular-recognition hybrids of natural polymers, in Polymers for Advanced

Technologies 13(10–12), 1019–1031 (2002).

95. D. Rousseau, Fat crystals and emulsion stability—A review, Food Res. Int. 33,

3–14 (2000).

96. D. Johansson, B. Bergnstahl, and E. Lundgren, Wetting of fat crystals by

triglycerides oil and water. I. The effect of additives, J. Am. Oil Chem. Soc. 72,

921–932 (1995).

97. D. Johansson, and B.Bergnstahl, Wetting of fat crystals by triglycerides oil

and water. II. Adhesion to the oil/water interface, J. Am. Oil Chem. Soc. 72,

933–938 (1995).

98. K. P. Oza, and S. G. Frank, Microcrystalline cellulose stabilized emulsions,

J. Dispersion Sci. Technol. 7, 543–561 (1986).

99. A. J. Khopade, and N. K. Jain, A stable multiple emulsion system bearing

isoniazid: Preparation and characterization, Drug Dev. Ind. Pharm. 24(3),

289–293 (1998).

100. N. Garti, A. Aserin, and I. Tiunova, Double emulsions of water-in-oil-in-

water stabilized by alpha-form fat microcrystals. Part I: Selection of

emulsifiers and fat microcrystalline particles, J. Am. Oil Chem. Soc. 76(3),

383–389 (1999).

101. T. Sekine, K. Yoshida, and F. Matsuzaki, A novel method for preparing oil-

in-water-in-oil type multiple emulsions using organophilic montmorillonite

clay mineral, J. Surfact. Detergents 2(3), 309–315 (1999).

102. P. Walstra, On the usefulness of dairy research, Int. Dairy J. 8, 155–161

(1998).


103. D. A. Edwards, and D. T. Wasan, A micromechanical model of linear surface

rheological behavior, Chem. Eng. Sci. 46, 1247–1257 (1991).

104. W. Kiosseoglou, Minor surface-active lipids of olive oil and viscoelasticity of

protein films at the olive oil-water interface, J. Dispersion Sci. Technol. 13,

135–144 (1992).

105. A. Vaziri, and B. Warburton, Improved stability of w/o/w multiple emulsion

by addition of hydrophilic colloid components in the aqueous phase,

J. Microencapsulation 12(1), 1–5 (1995).

106. S. Susuki, and J. K. Lim, Microencapsulation with carrageenan, locust bean

gum mixture in a multiphase emulsification technique for sustained drug-

release, J. Microencapsulation 11(2), 197–203 (1994).

107. D. Clausse, Thermal behavior of emulsions studied by differential scanning

calorimetry, J. Therm. Anal. Calorim. 51(1), 191–201 (1998).

108. D. Clausse, I. Pezron, and L. Komunjer, Stability of w/o and w/o/w emulsions

as a result of partial solidification, Colloids Surfaces A: Physicochemical and

Engineering Aspects 152, 23–29 (1999).

109. Y. Ogawa, Y. M. Yamamoto, H. Okada, H. T. Yashiki, and T. Shimamoto,

New technique to efficiently entrap leuprolidine acetate into microcapsules of

polylactic acid of poly(lactic/glycolic)acid, Chem. Pharm. Bull. 36, 1095–1103

(1988).

110. M. J. Blanco-Prieto, F. Delie, A. Fattal, A. Tartar, F. Puisieux, A. Gulik, and

P. Couvreur, Characterization of V3 BRU peptide loaded small PLGA

microspheres prepared by a W1/O/W2 emulsion solvent evaporation method,

Int. J. Pharm. 111, 137–145 (1994).

111. C. H. Schugens, N. Laruelle, N. Nihant, C. Grandfils, R. Jerome, and P.

Teyssie, Effect of the emulsion stability on the morphology and porosity of

semicrystalline poly-L-lactide microparticles prepared by w/o/w double

emulsion evaporation, J. Control. Release 32(2), 161–176 (1994).

112. N. Nihant, C. Schugens, C. Grandfils, R. Jerome, and P. Teyssie, Polylactide

microparticles prepared by double emulsion-evaporation. 2. Effect of

thepoly(lactide-co-glycolide) composition on the stability of the primary and

secondary emulsions, J. Colloid Interf. Sci. 173(1), 55–65 (1995).

113. M. J. Blanco-Prieto, E. Leo, F. Delie, A. Gulik, P. Couvreur, and E. Fattal,

Study of the influence of several stabilizing agents on the entrapment and in

vitro release of PBC from poly(lactide-co-glycolide) microspheres prepared by

a w/o/w solvent evaporation method, Pharm. Res. 13, 1137–1139 (1996).

114. R. Cortesi, F. Bortolotti, E. Menegatti, C. Nastruzzi, and E. Esposito,

Production and characterization of biodegradable microparticles for the

controlled delivery of proteinase inhibitors, Int. J. Pharm. 129, 263–273

(1996).

115. P. Couvreur, M. J. Blanco-Prieto, F. Puisieux, B. Roques, and E. Fattal,

Multiple emulsions technology for the design of microspheres containing

peptides and oligopeptides, Adv. Drug Delivery Rev. 28, 85–96 (1997).

116. M. J. Blanco-Prieto, E. Fattal, A. Gulik, B. Dedieu, and P. Couvreur,

Characterization and morphological analysis of a cholecystokinin derivative


peptide-loaded poly(lactide-co-glycolide)microspheres prepared by a w/o/w

emulsion solvent evaporation method, J. Control. Release 43, 81–87 (1997).

117. Y. Maa, and C. Hsu, Liquid–liquid emulsification by rotor/stator homo-

genisation, J. Control. Release 38, 219–228 (1996).

118. H. Rafati, A. Coombes, J. Adler, J. Holland, and S.S. Davis, Protein loaded

poly(DL-lactide-co-glycolide) microparticles for oral administration:

Formulation, structural and release characteristics, J. Control. Release 43,

89–102 (1997).

119. N. Ubrich, J. Ngondi, C. Rivat, M. Pfister, C. Vigneron, and P. Maincent,

Selective in vitro removal of anti-A antibodies by adsorption on encapsulated

erythrocytes–ghosts, J. Biomed. Mater. Res. 37, 155–160 (1997).

120. Y. F. Maa, and C. C. Hsu, Effect of primary emulsions on microsphere size

and protein-loading in the double emulsion process, J. Microencapsulation

14, 225–241 (1997).

121. E. Leo, S. Pecuet, J. Rojas, and P. Couvreur, Changing the pH of the external

aqueous phase may modulate protein entrapment and delivery from

poly(lactide-co-glycolide) microspheres prepared by a w/o/w evaporation

method, J. Microencapsulation 15, 421–430 (1998).

122. M. F. Zambaux, F. Bonneaux, R. Gref, E. Maincent, M. Dellacherie,

J. Alonso, P. Labrude, and C. Vigneron, Influence of experimental

parameters on the characteristics of poly(lactic acid) nanoparticles prepared

by double emulsion method, J. Control. Release 50, 31–40 (1998).

123. T. G. Park, M. J. Alonso, and R. Langer, Controlled release of proteins from

poly(L-lactic acid) coated polyisobutylcyanoacrylate microcapsules, J. Applied

Polymer Sci. 52, 1797–1807 (1994).

124. B. De Cindio, and D. Cacace, Formulation and rheological characterization

of reduced-calorie food emulsions, Int. J. Food Sci. Technol. 30, 505–514 (1995).

125. B. De Cindio, G. Grasso, and D. Cacace, Water-in-oil-in-water emulsions for

food applications, Food Hydrocolloids 4, 339–353 (1991).

126. D. K. Asahi, Double emulsified oil-fat composition comprises outer phase of

oil and fat inner phase of oil-in-water emulsified fats, and oil and fat phase

containing mustard oils, Japanese patent 87016619, 1987.

127. K. K. Yoshihara Seiyu, Oil and fat cooking product containing double

emulsion of liquid oil and fat of above 30�C melting points, Japanese patent

60091939, 1985.

128. K. Suzuki, K. Uehara, and H. Omura, Oil-in-water-in-oil double emulsified

fat or oil composition for cream, spread confectionery or baking use,

European patent 425958, 1991.

129. S. Tsukishima, Double emulsion type filling cream obtained by adding fat or

oil to aqueous phase containing saccharides, protein and/or emulsifiers and

mixing with plastic fat or oil, Japanese patent 3290152, 1991.

130. C. Laugel, P. Rafidison, G. Potard, L. Aguadisch, and A. Baillet, Modulated

release of triterpenic compounds from a o/w/o multiple emulsion formulated

with dimethicones: Infrared spectrophotometric and differential calorimetric

approaches, J. Control. Release 63(1–2), 7–17 (2000).


131. V. Muguet, M. Seiller, G. Barratt, D. Clausse, J.P. Marty, and J.L. Grossiord,

w/o/w multiple emulsions submitted to a linear shear flow: Correlation

between fragmentation and release, J. Colloid Interf. Sci. 218, 335–337 (1999).

132. S. Geiger, N. Jager-Lezer, S. Tokgoz, M. Seiller, and J.L. Grossiord,

Characterization of the mechanical properties of a water/oil/water multiple

emulsion oily membrane by a micropipette aspiration technique, Colloids

Surfaces A: Physicochemical and Engineering Aspects, 157(1–3), 325–332

(1999).

133. J. L. Grossiord, and M. Seiller, Rheology of w/o/w multiple emulsions:

formulation, characterization and breakup mechanisms, in Multiple

Emulsions: Structure, Properties and Applications (J.L. Grossiord and M.

Seiller, eds.), Editions de sante, Paris, 1998, pp. 169–192.

134. P. Stroeve, and P. P. Varanassi, An experimental study on double emulsion

drop breakup in uniform shear flow, J. Colloid Interf. Sci. 99, 360–373 (1984).

135. R. Pal, Multiple O/W/O emulsion rheology, Langmuir 12(9), 2220–2225

(1996).

136. Y. Kawashima, T. Hino, H. Takeushi, T. Niwa, and K. Horibe, Rheological

study of w/o/w emulsion by a cone-and-plate viscometer: negative thixotropy

and shear-induced phase inversion, Int. J. Pharm. 72, 65–77 (1991).


11Structure, Mechanics, and Rheologyof Concentrated Emulsions andFluid Foams

H. M. PrincenConsultant, Flemington, New Jersey, U.S.A.

I. BACKGROUND

Whether enjoying the luxury of a bubble bath or enduring the drudgeryof washing dishes, one is likely to be struck by the beauty and intricatestructure of foams, froths, or ‘‘suds.’’ Keen observers may even notice theunusual elastic and yield properties, not seen in the constituent aqueous andgaseous phases. Scientifically, the interest in and the study of foams have beentruly multidisciplinary and have not been confined to chemists, engineers,and physicists. Foams have traditionally inspired mathematicians fortheir geometric properties and as equilibrium structures in which the surfacearea is minimized (1). Metallurgists (2) have realized the similarity betweenfoams and polycrystalline metals, both in their structure and coarseningbehavior (grain growth). Similarly, botanists and life scientists in generalhave noticed strong structural parallels between foams and living tissues (3).

Gas–liquid foams are abundant in nature and their technologicalapplications are numerous. They are used to advantage in fire fighting,enhanced oil recovery, foods (e.g., whipped cream), cosmetics (e.g., shavingcream), and in many other ways. The ‘‘world of foams’’ may be considerablyexpanded by the realization that concentrated liquid–liquid emulsions,although generally characterized by a much smaller mean size of thedispersed units, are structurally identical to gas–liquid foams, which isreadily revealed under the microscope. Macroscopically, they behave like


viscoelastic gels, mayonnaise being a good example. Such emulsions havebeen variously referred to as high-internal-phase-ratio emulsions (HIPREs),biliquid foams, ‘‘aphrons,’’ or, simply, highly concentrated emulsions.Although they lack the compressibility of gas–liquid foams, they behavesimilarly in all other respects. Detailed study of such emulsion systemsstarted rather recently and may perhaps be traced to the attempts ofLissant (4–6) and Beerbower et al. (7–10) to design safer aviation androcket fuels, in which fuel droplets are tightly packed inside a continuousaqueous phase. Reverse (i.e., concentrated water-in-oil) systems can bereadily prepared as well. They find application in the high-explosives area,but have particular appeal in the foods and cosmetics industries. Whatentrepreneur’s mouth would not water at the prospect of being able tosell a product that is at least 90% water and yet is luxuriously rich andcreamy? Lissant, in particular, patented numerous potential applications inthese areas (e.g., Ref. 11). In yet other applications, the oil phase, eitherexternal or internal, can consist of a polymerizable monomer. Subsequentpolymerization by heat or radiation can lead to interesting polymers orstructurally unique materials (e.g., Refs. 6 and 12–16).

Because of all of these scientific and technological aspects, a thoroughunderstanding of foams and concentrated emulsions is highly desirable. Inresponse to this need, there has lately been a clear upsurge in interest, againfrom a variety of disciplines, and considerable progress has been and isbeing made. Several comprehensive textbooks on emulsions and foamshave recently been published (17–22). We believe that the overlap withthis review is minimal.

II. INTRODUCTION

In general, when a fluid phase (liquid or gas) is dispersed in an immiscibleliquid to form drops or bubbles, there is a tendency for the phases toseparate again to reduce the augmented surface free energy. With purephases, this proceeds by rapid coalescence of approaching dispersed entities,as there is no barrier against rupture of the intervening liquid film. Stabilityor, more correctly, metastability can be conferred by adsorption of surfac-tants, polymers, or finely divided solid particles at the interface. By thisexpedient, coalescence can often be suppressed completely. However, thiswill not prevent ultimate phase separation, as there is another mechanismfor reducing the surface area, namely Ostwald ripening. By this mechanism,large bubbles or drops grow at the expense of small ones by dissolution anddiffusion of the dispersed phase in response to the higher Laplace pressure inthe latter ones. Because gases tend to have greater solubility and diffusivity


in a given continuous liquid than do most other liquids, this process isgenerally much more rapid in foams than in emulsions. Indeed, whereasmost foams will not survive for more than a few hours—even in the absenceof coalescence—it is relatively easy to prepare concentrated emulsionswhose drop size distribution does not change perceptibly for monthsor years. They are kinetically or operationally (although not thermodyna-mically) stable. For this and many other reasons, emulsions may be bettercharacterized and their properties more reliably investigated experimentallythan is possible with foams. Thus, to learn about foam behavior throughexperiments, we recommend one look at concentrated emulsions instead. Inthe same vein, we may use the terms ‘‘bubble’’ and ‘‘drop’’ interchangeably.

In this review, we will only consider stable dispersions, in which coa-lescence has been totally suppressed. We further restrict ourselves to highlyconcentrated dispersions, in which the volume fraction of the dispersedphase, �, exceeds a critical value �0 where the properties start to changedrastically. This critical volume fraction corresponds to that of a system ofclose-packed spheres having the same drop volume distribution as the dis-persion. The term ‘‘close packed’’ is somewhat ambiguous and the corre-sponding volume fraction is not always clearly defined and/or established.Although monodisperse spheres can, in principle, be packed to a maximumdensity of �0¼ 0.7405, this value is rarely achieved. In practice, one is morelikely to achieve only random close packing, which is considerably less dense(�0� 0.64) due to the voids created by ‘‘arching.’’ There is a persistent myththat the packing density of a polydisperse system is characterized by�0>0.7405. It is true that the voids in a close-packed system of spherescan be filled sequentially with smaller and smaller spheres of very specificsizes until �0� 1. However, this would require a unique multimodal sizedistribution as well as a unique spatial distribution, neither of which arelikely to be ever encountered in practice. It is our experience with typical,unimodal polydisperse emulsions that the spherical droplets arrangethemselves at a packing density that, although considerably larger thanthe 0.64 expected for the random close-packed monodisperse case, is closeto but slightly smaller than 0.74. Although the actual value mustdepend somewhat on the details of the size distribution, we estimate that0.70< �0<0.74 in most practical cases, where conventional means are usedto prepare the emulsion (23,24).

There are reasons why the effective value of �, including that of �0,may deviate from the apparent value. If the thickness, h, of the stabilizedfilm of continuous phase, separating the dispersed drops or bubbles, is notinsignificant compared to the drop or bubble radius, R, then the effectivevolume of each drop must be augmented by that of a surrounding sheath ofthickness h/2. This leads to a somewhat larger effective volume fraction,


�e, which is given (23) by

�e ¼ � 1��

�0

� �1=3h

2R

" #�3

� � 1þ3:15h

2R

� �

ð1Þ

The latter form is a good approximation for any �>�0 and h/R<<1. Inmost foams, the effect is expected to be minimal, as the bubbles tend to berelatively large. For emulsions of small drop size, however, the effect may beconsiderable and the peculiar properties resulting from extreme crowdingmay commence at an apparent volume fraction that is considerably smallerthan one would expect for zero film thickness. For example, in an emulsionwith droplets of 2R¼ 1 mm and h¼ 50 nm, the effective volume fractionalready reaches a value of 0.74 at an apparent volume fraction of onlyabout 0.64! The finite film thickness may, for example, result from electro-static double-layer forces (25) or adsorbed polymers. In what follows, weshall assume zero film thickness, with the understanding that Eq. (1) is to beinvoked whenever h/R 6¼ 0.

Another complication arises when strong attractive forces operatebetween the drops or bubbles. This may lead to a finite contact angle, �,between the intervening film (of reduced tension) and the adjacent bulk inter-faces (23,26–28). Under those conditions, droplets will spontaneously deforminto truncated spheres upon contact and can thus pack to much higher den-sities. For monodisperse drops, the ideal close-packed density, consistentwith minimization of the system’s surface free energy, is given (23) by

�0ð�Þ ¼ 0:7405 �5

cos3 �þ

9

cos2 �� 3

� �

ð2Þ

which is valid up to �¼ 30�, where �0¼ 0.964. For �¼ 0, we recover�0¼ 0.7405, and �0 is expected to reach unity when � exceeds 35.26�

(23,28). In the latter limit, all of the continuous phase (except that in theintervening films) should, in principle, be squeezed out spontaneously. Inpractice, however, one tends to find just the opposite; that is, when � islarge, the droplets spontaneously flocculate into a rather open structure inwhich �0<0.7405. The situation is similar to that of a flocculated soliddispersion whose sediment volume is generally greater than that of a stabledispersion. Apparently, the strong attractive forces prevent the droplets fromsliding into their energetically most favorable positions, leaving large voids inthe otherwise dense structure. Nevertheless, the structure may be irreversiblydensified to approach the condition prescribed by Eq. (2) by centrifugationand subsequent relaxation (23,27). Foams and emulsions in which � 6¼ 0 haveonly been studied occasionally and will rarely be touched upon in this review.


III. STRUCTURAL ELEMENTS

As discussed earlier, the nature and properties of fluid–fluid dispersionsstart to change drastically when the volume fraction approaches or exceeds�0. A certain rigidity sets in, because the drops or bubbles can no longermove freely past each other.

As the volume fraction is raised beyond �0, the drops lose their spheri-city and are increasingly deformed while remaining separated by thin stablefilms of continuous phase. At sufficiently high �, the drops become dis-tinctly polyhedral, albeit with rounded edges and corners. At this stage,the continuous phase is confined to two structural elements: linear Plateauborders with an essentially constant cross section over some finite length,and tetrahedral Plateau borders, where four linear borders converge (Fig. 1a).

Each linear border is generally curvilinear and fills the gap between therounded edges of three adjoining polyhedral drops. In cross section, its sidesare formed by three arcs, each pair of which meets tangentially to form thethin film separating the corresponding droplet pair (Fig. 1b). The pressures inthe drops are related to the mean curvatures of the intervening films through

p1 � p2 ¼ 2C12

p2 � p3 ¼ 2C23

p3 � p1 ¼ 2C31

ð3Þ

where is the interfacial tension between the continuous and dispersedphases and the sign of each film curvature C is taken positive (negative)

Figure 1 (a) Four linear Plateau borders meeting in a tetrahedral plateau border;

(b) cross section through a linear Plateau border and its three associated films and

drops.


if the pressure in the drop indicated by the first index is the higher (lower)one. Adding Eqs. (3) leads to the following relationship between the threemean film curvatures:

C12 þ C23 þ C31 ¼ 0 ð4Þ

The pressure inside the linear Plateau border, pb, is given by

pb ¼ p1 � c1 ¼ p2 � c2 ¼ p3 � c3 ð5Þ

where c1, c2, and c3 are the curvatures of the border walls and are all countedas positive. Since all Plateau borders are connected, they are in hydrostaticequilibrium.

Normally, an ambient gaseous atmosphere of pressure P surrounds thedispersion. Relative to this ambient pressure, pb is lower and given (29) by

pb ¼ P� c Ctj j ð6Þ

where Ctj j is the absolute value of the curvature of the free continuous-phasesurface at the dispersion–atmosphere boundary (i.e., between the exposedbubbles) and c is the surface tension of the continuous phase [c¼ forfoams, but c 6¼ for emulsions (29) unless the ‘‘ambient atmosphere’’ con-sists of bulk dispersed liquid].

The excess pressures in the drops, relative to that in the interstitialcontinuous phase, pb, are often referred to as their capillary pressures, pc.For example,

ð pcÞ1 ¼ p1 � pb ¼ c1 ð7Þ

It is clear that, in general, the capillary pressure varies from drop to drop.When Eqs. (3) are combined with Eqs. (5), the following relationships

between the curvatures of the films and those of the Plateau border walls areobtained:

2C12 ¼ c1 � c2

2C23 ¼ c2 � c3

2C31 ¼ c3 � c1

ð8Þ

For each film to be stable, it must be able to develop an internal,repulsive disjoining pressure �d to counteract the capillary suction actingat the film–Plateau border junction. At equilibrium, it can be readily shown


from the above that

ð�dÞ12 ¼ðc1 þ c2Þ

2


2


2

ð9Þ

Thus, the disjoining pressures in three confluent films are, in general,unequal. It turns out that the difference in the disjoining pressures in twoof the films is defined by the curvature of the third film. For example,from Eqs. (9) and (8),

ð�d Þ31 � ð�dÞ23 ¼ðc1 � c2Þ

2¼ C12 ð10Þ

[The inequality of the disjoining pressures implies that the films may haveslightly different equilibrium thicknesses and tensions. In extreme cases (30),this may lead to sensible deviations from Plateau’s first law of foamstructure, stated below.]

As the volume fraction approaches unity, the linear Plateau bordershrinks into a line. In this ‘‘dry-foam’’ limit, mechanical equilibriumdemands that the three films—of presumed equal tensions—meet pairwiseat angles of 120� along this line (Plateau’s first law of foam structure).However, even when the Plateau border is finite and the films do not reallyintersect, the principle may well hold when applied to the virtual line ofintersection that is obtained when the films, while maintaining their curva-tures, are extrapolated into the border (dashed lines in Fig. 1b). A rigorousproof has been published by Bolton and Weaire (31) for two-dimensional(2D) foams, in which the Plateau borders are rectilinear. To our knowledge,no proof has yet been presented for the more general case of curvilinearborders in three-dimensional (3D) space. In fact, because the Plateau bordercan be viewed as a line with a line tension (32), this broader statement ofPlateau’s first law may not strictly apply when the border has some finitelongitudinal curvature.

A tetrahedral Plateau border is formed by the confluence of fourlinear Plateau borders (Fig. 1a). It fills the gap between the rounded cornersof four adjoining polyhedral drops. The pressure in the tetrahedral borderis, of course, equal to that in each of the outgoing linear borders, which setsthe curvature of each of its four bounding walls. In the dry-foam limit(�! 1), the tetrahedral border reduces to a point (‘‘vertex’’ or ‘‘node’’),


where the four linear borders meet pairwise at the angle of cos�1(�1/3)¼109.47� (Plateau’s second law of foam structure). The principle probablyremains valid for finite borders, when applied to the point where the fourvirtual lines of film intersection (see Fig. 1) meet upon extension into thetetrahedral border.

IV. OVERALL STRUCTURE AND OSMOTIC PRESSURE

Having described the structural elements of foams approaching the dry-foam limit (�!1), it is still a daunting task to describe the structure andproperties of the system as a whole. The task is even more difficult forsystems in which �0 is exceeded, but the polyhedral regime has not yet beenreached. In this case, the drops have exceedingly complex shapes, and linearand tetrahedral Plateau borders, as defined earlier, are not present. Muchcan be learned about the qualitative behavior by considering 2D modelsystems, in which the drops do not start out as spheres but as parallelcircular cylinders and tetrahedral plateau borders do not arise. We shallfirst consider the particularly simple, perfectly ordered monodisperse case,with a subsequent gradual increase in complexity.

[Lest the reader think that 2D foams are just figments of the imagina-tion, it must be pointed out that they can be generated—or at least closelyapproximated—by squeezing a 3D foam between two narrowly spaced,wetted, transparent plates (2,33–37). Structurally, even closer realizationsmay be obtained in phase-coexistence regions of insoluble monolayers ofsurface-active molecules at the air–water interface (38), where the role ofsurface tension is taken over by the line tension at the phase boundaries.]

A. Monodisperse, Perfectly Ordered 2D System

Such a system has been discussed in detail in Ref. 39. In the absenceof gravity, the circular cylinders of radius R arrange in hexagonal packing(Fig. 2a) at a volume fraction �0¼�/2

ffiffiffi3

p¼ 0.9069. In cross section, each

circular drop can be thought to be contained within a regular hexagon ofside length a0¼ 2R/

ffiffiffi3

p. As the volume fraction is increased, the drop is

flattened against its six neighbors to form a hexagon of side length a(<a0) but with rounded corners described by circular arcs of radius r(Fig. 2b). At constant drop volume, one finds

r

R¼

�01� �0

� �1=21� �

�

� �1=2

ð11Þ


The capillary pressure in each drop is given by pc¼ /r or, when scaled bythe initial capillary pressure ( pc)0¼ /R, by

~ppc ¼ pc=ð pcÞ0 ¼ R=r ð12Þ

In the above process, the surface area of each drop, per unit of length,increases from S0¼ 2�R to S¼ 6(a� 2r/

ffiffiffi3

p)þ 2�r, which, at constant drop

volume, can be shown to lead to

S

S0¼

1

ð�0�Þ1=2

1� ð1� �0Þ1=2

ð1� �Þ1=2� �

ð13Þ

This function has been plotted in Fig. 3. In the limit of �¼ 1, the scaled surfacearea reaches a maximum that is given by

S1

S0¼

1

�1=20

¼ 1:0501 ð14Þ

The scaled surface area and its variation with � are of crucial impor-tance in the definition and evaluation of the osmotic pressure, �, of a foamor emulsion. We introduced the concept in Ref. 39, where it was referred toas the ‘‘compressive pressure,’’ P. It has turned out to be an extremelyfruitful concept (24,29,40). The term ‘‘osmotic’’ was chosen, with somehesitation, because of the operational similarity with the more familiarusage in solutions. In foams and emulsions, the role of the solute moleculesis played by the drops or bubbles; that of the solvent is played by thecontinuous phase, although it must be remembered that the nature of theinteractions is entirely different. Thus, the osmotic pressure is defined asthe pressure that needs to be applied to a semipermeable, freely movable

Figure 2 (a) Uncompressed cylindrical drops in hexagonal close packing

(�¼�0¼ 0.9069); (b) compressed drop (0.9069<�<1).


membrane, separating a fluid–fluid dispersion from its continuous phase,to prevent the latter from entering the former and to thereby reduce theaugmented surface free energy (Fig. 4). The membrane is permeable to all ofthe components of the continuous phase but not to the drops or bubbles.As we wish to postpone discussion of compressibility effects in foams untillater, we assume that the total volume (and therefore the volume of thedispersed phase) is held constant.

Figure 3 Scaled surface area, S/S0, for monodisperse 2D drops as a function of

volume fraction.

Figure 4 Semipermeable membrane separating dispersion from continuous phase;

pressure to prevent additional continuous phase from entering the dispersion is the

‘‘osmotic’’ pressure,�. [FromRef. 40. Copyright (1986) American Chemical Society.]


As long as the membrane is located high up in the box in Fig. 4,the emulsion or foam may be characterized by �< �0 and �¼ 0. As themembrane moves down, a point is reached where �¼�0. Any further down-ward movement requires work against a finite pressure �, reflecting theincrease in the total surface area as the drops are deformed; that is,

�� dV ¼ �� dV2 ¼ dS ðconstant V1Þ ð15Þ

where V is the dispersion volume, V1 is the volume of the dispersed phase,V2 is the volume of the continuous phase in the dispersion, and is assumedto be constant. Because V¼V1þV2 and �¼V1/V, Eq. (15) leads to thecompletely general expression

� ¼ �2dðS=V1Þ

d�¼ �2

S0

V1

dðS=S0Þ

d�ð16Þ

where S/V1 is the surface area per unit volume of the dispersed phase.Alternatively, as shown in Ref. 29, � may be equated to the pressuredifference between an ambient atmosphere and the continuous phase inthe dispersion, or from Eq. (6):

� ¼ P� pb ¼ c Ctj j ð17Þ

For yet a third useful way to express �, see Refs. 24, 29, and 40.For the special case of a monodisperse 2D system,

S0

V1¼

2

Rð18Þ

which, when combined with Eqs. (16) and (13), results in

� ¼

R

�

�0

� �1=21� �01� �

� �1=2

�1

" #

ð19Þ

or, in reduced form,

~��

=R¼

�

ð pcÞ0¼

�

�0

� �1=21� �01� �

� �1=2

�1

" #

ð20Þ

where �0¼ 0.9069. Figure 5 shows the dependence of ~�� on �.The suggestion has been made (Exerowa, personal communication,

1990), since withdrawn (20,41), that � and pc are really identical. It is


clear from the above that this is not so. In fact, examination of Eqs. (20),(11), and (12) shows that, at least for this simple model system,

~ppc � ~�� ¼�

�0

� �1=2

ð21Þ

At �¼�0,

~ppc � ~�� ¼ 1 ð22Þ

At the upper limit of �¼ 1,

~ppc � ~�� ¼ ��1=20 ¼

S1

S0¼ 1:0501 ð23Þ

Both ~ppc and ~�� tend to infinity in this limit, but the relative differencebetween them tends to zero. This is the regime of concern in much ofthe interesting work of Exerowa et al. (e.g., Refs. 42 and 43), wherethe difference between the capillary and osmotic pressures may there-fore, indeed, be safely ignored (41). However, this is not so in generaland we shall demonstrate in Sect. V that � is a much more usefuland informative parameter than pc.

Before leaving this topic, it should be mentioned that modifications ofmost of the above expressions have been derived to take account of finitefilm thickness, finite contact angle at the film–Plateau border junction, or

Figure 5 Reduced osmotic pressure as a function of � for a perfectly ordered 2D

system.


both (39). Finally, it must be realized that a monodisperse 2D system doesnot necessarily pack in the perfectly ordered, hexagonal state depicted inFig. 2. Herdtle et al. (personal communication, 1993) have constructedhighly disordered, yet monodisperse 2D dry foams with periodic boundaries(Fig. 6), in which all films meet at angles of 120� and all film curvaturessatisfy Eq. (4). In a sense, these systems are monodisperse only in avolumetric sense, but are topologically disordered or ‘‘polydisperse.’’ They areequilibrium structures, whose surface energy, although at a local minimum,must be higher than that of the perfectly ordered hexagonal system. Becausethe bubble pressures are not the same, such a system is bound to coarsen,thereby reducing its total surface energy. In practice, disorder of this typemay be imposed by the finiteness of any system with bounding walls. If thewalls are wetted by the continuous phase, then the outer films must bedirected normal to the walls, which is generally incompatible with a perfectlyordered internal structure. As we shall see, this complication arises in 3Dfoams as well.

Figure 6 Topologically disordered, but volumetrically monodisperse 2D system

(�¼ 1) with periodic boundaries; each shade corresponds to drops with a certain

number of sides (e.g., the unshaded drops all have six sides). (Courtesy of T. Herdtle

and A. M. Kraynik.)


B. Polydisperse 2D Systems

In the last decade or so, much progress has been made toward a morecomplete understanding of these disordered structures. Most work relies onthe computer generation of disordered (volumetrically) polydisperse struc-tures with periodic boundary conditions, in which the film angles and curva-tures obey the rules set forth in Sect. III. For a recent review, see Ref. 33. Anexample, taken from Ref. 44, is shown in Fig. 7. The structure contains manybubbles that are not hexagons, but it is readily proven that the averagenumber of sides is still six (44). Simpler and very special types of (volumetric)polydispersity and disorder have been considered by Khan and Armstrong(45) andKraynik et al. (46). In these cases, illustrated in Fig. 8, the topology ispreserved. All bubbles are still hexagons and all films remain flat; the bubbles,therefore, do not coarsen with time. The first system (Fig. 8a) is simplybimodal and is obtained by increasing or decreasing the height of all bubblesin a given row. The second system (Fig. 8b) is much more disordered and canbe generated from the monodisperse system by randomly increasing (ordecreasing) each bubble area, as illustrated in Fig. 9, with the limitationthat no vertices ever touch or cross over, lest Plateau’s first law be violatedand resultant (so-called T1) rearrangements lead to a much more complexstructure. The total surface area is not affected by such transformations, sothat, as in the monodisperse case,

S1

S0¼ 1:0501 ð24Þ

Figure 7 Computer-generated, topologically disordered and volumetrically

polydisperse 2D system (�¼ 1) with periodic boundaries. (From Ref. 44, with

permission from Taylor & Francis Ltd.)


This is not necessarily true for the more general structures such as thatin Fig. 7, which is both volumetrically and topologically polydisperse.Unfortunately, although presumably available as a result of the numericalsimulations, the value of S1/S0 and how it varies with the details of the sizedistribution appear not to have been reported for these cases.

Starting from a dry-foam system as in Fig. 7, the volume fraction canbe lowered by ‘‘decorating’’ each vertex with a Plateau border, whose wallcurvatures obey the rules set forth in Sect. III (31). As the volume fraction islowered by increasing the size of the Plateau borders, a point is soon reachedwhere adjacent Plateau borders ‘‘touch’’ and subsequently merge into singlefour-sided borders. Bolton and Weaire (47) have followed this process down

Figure 8 (a) Simplest case of volumetrically bimodal 2D system; (b) more highly

disordered, volumetrically polydisperse hexagonal 2D system (�¼ 1). The cluster

of darkly outlined drops forms the repeating unit. (Courtesy of A. M. Kraynik.

Similar structures appear in Ref. 46.) In both cases, the system is topologically

‘‘monodisperse.’’

Figure 9 Recipe for creating a volumetrically polydisperse (but topologically

monodisperse) hexagonal system from perfectly ordered 2D system; total surface

energy remains unchanged.


to the volume fraction �c, where all bubbles are spherical and structuralrigidity is lost. This is perhaps the most satisfactory definition of �0.Their finding suggests that, for that particular system, �c equaled 0.84(not 0.9069), which happens to be close to the random packing density of(monodisperse) circular disks. Using similar computer simulations, Hutzlerand Weaire (48) calculated the osmotic pressure and found it to obeyEq. (19) closely in the ‘‘drier’’ regime. It started to deviate at lowervolume fractions and did not reach zero until � dropped to about 0.82,which is close to � at the rigidity loss transition.

C. Monodisperse 3D Systems

Ideally, uniform spheres arrange in ‘‘hexagonal close’’ packing, which isface-centered cubic (fcc), at �0¼�

ffiffiffi2

p/6¼ 0.7405. The role of the circum-

scribing hexagon in monodisperse 2D systems is taken over by the rhombicdodecahedron (Fig. 10). As the volume fraction is raised, each drop flattensagainst its 12 neighbors. This process has been described by Lissant (4,5),who considered the drop to be transformed into a truncated sphere and each

Figure 10 Spheres in hexagonal close packing (fcc), each occupying a rhombic

dodecahedron. (From Ref. 4, with permission from Academic Press.)


film to be circular, at least until it reaches the sides of the diamond faces(Fig. 11). This is incompatible with a zero contact angle at the film edge.Moreover, at constant drop volume, this model would imply decreasingcapillary (and osmotic) pressure with increasing �, which is clearly incon-sistent. In reality, the problem is much more complicated; the drop cannotremain spherical and the films must be noncircular. Using Brakke’s nowfamous ‘‘Surface Evolver’’ computer software (49), Kraynik and Reinelt(50), and Lacasse et al. (51) have correctly and accurately solved thisproblem for this and other structures (discussed later in this subsection).

As suggested already by Lissant (4,5), the packing is likely to changeabove some critical value of �. It is clear that if the dodecahedral packingwere to persist up to �¼ 1, Plateau’s second law would be violated at 6 ofthe 14 corners of the polyhedron, because 8 linear borders would convergethere, rather than the mandatory 4. Lissant proposed that the structurechanges to a body-centered cubic (bcc) packing of planar tetrakaidecahedra(truncated octahedra; see Fig. 12a). However, such a structure satisfiesneither of Plateau’s laws. In this dry-foam regime, Kelvin’s ‘‘minimal

Figure 11 Each drop flattens against its neighbors as the volume fraction

increases; a stable thin film of continuous phase separates neighboring drops. (From

Ref. 4, with permission from Academic Press.)


tetrakaidecahedron’’ (Fig. 12b), which is obtained by slight distortion of itsplanar counterpart, solves this problem and has long been considered as themost satisfactory candidate for the drop shape. It has 6 planar quadrilateralfaces, 8 nonplanar hexagonal faces of zero mean curvature, and 36 identicalcurved edges. In a space-filling ensemble of such polyhedra, Plateau’s firstand second laws are fully satisfied. Kelvin derived approximate expressionsfor the shape of the hexagons and the sides (52–54). Based on that model,Princen and Levinson (55) calculated the length of the sides and the surfaceareas of the quadrilateral and hexagonal faces, relative to those of the parentplanar tetrakaidecahedron of the same volume. They arrived at the follow-ing result for the increase in surface area as a spherical drop transforms intoa Kelvin tetrakaidecahedron of the same volume:

S1

S0¼ 1:0970 ð25Þ

(This compares to values of 1.0990 for the planar tetrakaidecahedron, 1.1053for the rhombic dodecahedron, and 1.0984 for the regular pentagonaldodecahedron. The latter—although often considered as a unit cell in foammodeling—is not really a viable candidate either, as it not only violatesPlateau’s laws but is also not space filling.)

More recently, Reinelt and Kraynik (56) have carried out more exactnumerical calculations on theKelvin cell, leading to the slightly higher value of

S1

S0¼ 1:0972 ð26Þ

Kelvin’s polyhedron would indeed represent the ideal drop shape in thedry-foam limit by effecting, in Kelvin’s own words, ‘‘a division of space with

Figure 12 (a) Planar tetrakaidecahedron (or truncated octahedron); (b) Kelvin’s

minimal tetrakaidecahedron (bcc).


minimum partitional area,’’ if he had added the proviso that this division is tobe accomplished with identical cells. It has recently been proven by Weaireand Phelan (57) that at least one structure of even lower energy exists, if thisrestriction is lifted. The Weaire–Phelan structure (Fig. 13), whose surfacearea is about 0.34% lower than that of Kelvin (i.e., S1/S0¼ 1.0936), hasrepeating units that contain 8 equal-volume cells: 2 identical pentagonaldodecahedra and 6 identical tetrakaidecahedra that each have 12 pentagonaland 2 hexagonal faces. The pressure in the dodecahedra is slightly higher thanthat in the tetrakaidecahedra. Perhaps surprisingly, neither the Kelvin northe Weaire–Phelan structure is rarely, if ever, encountered in actualmonodisperse foams (3). The reason for this may lie in small deviationsfrommonodispersity or, more likely, in the disturbing effects of the containerwalls, as alluded to already in connection with 2D foams. Alternatively, as thecontinuous phase is removed from between the initially spherical drops in fccpacking, slight irregularities in this drainage process may force the system toget trapped in a less-ordered structure that may be at a local surfacearea minimum but is separated from the lower-energy Kelvin and Weaire–Phelan structures by a significant barrier [cf. the difficulty one encountersin trying to build a 15-bubble cluster that has a Kelvin polyhedron at itscenter (58)].

Figure 13 Unit cell in Weaire–Phelan structure, containing 2 pentagonal

dodecahedra and 6 tetrakaidecahedra, each having 12 pentagonal and 2 hexagonal

faces. (Courtesy of A. M. Kraynik.)


Kraynik and Reinelt (50) and Lacasse et al. (51) have accurately com-puted the changes in surface area as a drop transforms from a sphere into aregular dodecahedron (fcc) or a Kelvin cell (bcc) with increasing volumefraction while maintaining zero contact angle. Expressed in terms of S/S0,the results are shown in Fig. 14. The Kelvin structure is internally unstablebelow �� 0.87. The results further indicate that the Kelvin cell becomes themore stable structure above �� 0.93. Also indicated is the limiting law for�! 1 for the dodecahedron. In that regime, linear Plateau borders ofconstant cross section run along the edges of the polyhedron. Their volumesand surface areas can be evaluated as a function of �, whereas the volumesand surface areas of the tetrahedral borders become negligible. For therhombic dodecahedron (24), this leads to

S

S0¼ 0:0686 1� 1:892 1� �1=2

� �� 3þ1:0367 ð� ! 1Þ

Kraynik and Reinelt (50) also evaluated the all-important osmoticpressure �(�), which, for 3D structures, is given by [cf. Eq. (16)]

� ¼ �2S0

V1

dðS=S0Þ

d�¼

3�2

R

dðS=S0Þ

d�

Figure 14 Scaled surface areas as a function of volume fraction for the rhombic

dodecahedral (fcc) and Kelvin structures (bcc). (From data kindly provided by A. M.

Kraynik and D. A. Reinelt.)


where R is the radius of the initially spherical drops, or

~��

=R¼ 3�2

dðS=S0Þ

d�ð27Þ

For the dodecahedron, the appropriate limiting law for �! 1 is given(24) by

~�� ¼�

=R¼ 0:5842�1=3

1� 1:892ð1� �Þ1=2� �2

ð1� �Þ1=2ð�! 1Þ ð28Þ

Figure 15 shows ~��ð�Þ for the dodecahedron and Kelvin cell.Detailed numerical calculations have been carried out by Bohlen et al.

(59) for the transition of monodisperse spheres in simple cubic packing(�0¼ 0.5236) to cubes (�¼ 1), for both zero and finite contact angles.Unfortunately, although the results are interesting, this kind of packing isnot realistic for foams and emulsions and will not be discussed further.

D. Polydisperse 3D Systems

This is, of course, the system of greatest interest from a practical point ofview. The detailed structure is exceedingly complex. As mentioned earlier,

Figure 15 Reduced osmotic pressure as a function of volume fraction for the

rhombic dodecahedral and Kelvin structures. (From data kindly provided by A. M.

Kraynik and D. A. Reinelt.)


even the value of �0 is not precisely defined and is expected to dependsomewhat on the details of the size distribution. Nevertheless, there isclear experimental evidence (23,24) that �0 is close to—or slightly smallerthan—0.7405 for ‘‘typical,’’ polydisperse, unimodal emulsions.

In the dry-foam limit, each polyhedral drop must satisfy Euler’sformula; that is,

v� eþ f ¼ 2 ð29Þ

where v is the number of vertices, e is the number of edges, and f is thenumber of faces. For an infinite number of space-filling polyhedra that aresubject to Plateau’s rules, a number of statistical relationships can bederived from Eq. (29) (60–62). Perhaps the most interesting of these is

h f i ¼12

6� heið30Þ

where h f i is the average number of faces per cell and hei is theaverage number of edges per face. Equation (30) is consistent with whatis expected for a monodisperse ‘‘Kelvin foam,’’ where h f i ¼ f¼ 14 andeh i¼ (6� 4þ 8� 6)/14¼ 5.143, or a Weaire–Phelan structure, where h f i¼(2� 12þ 6� 14)/8¼ 13.5 and eh i ¼ [2� 12� 5þ 6� (12� 5þ 2� 6)]/108¼5.111. As mentioned earlier, Matzke (3) found that, in a real, supposedlymonodisperse foam, Kelvin’s polyhedra did not occur and that pentagonalfaces were predominant. He found that h f i ¼ 13.70 and eh i ¼ 5.124, which is,again, consistent with Eq. (30). For a real polydisperse dry foam, Monnereauand Vignes-Adler (63) found h f i¼ 13.39 0.05 and eh i¼ 5.11, again in closeagreement with Eq. (30). These authors did not encounter any Kelvin cell (orWeaire–Phelan structure) either.

For �0<�<1, the drops go through a complex transition fromspheres to pure polyhedra. In this most general system, the osmotic pressureis given by

�ð�Þ ¼ �2S0

V1

dðS=S0Þ

d�¼

3�2

R32

dðS=S0Þ

d�ð31Þ

where R32 is the surface/volume or Sauter mean radius of the initially spheri-cal drops:

R32 �

PniR

3iP

niR2i

¼3V1

S0ð32Þ


Although R32 can be readily measured for any practical system, the complexgeometry does not allow the evaluation of S(�)/S0 and �(�) from firstprinciples. Instead, in the next section, we shall show how these and otherimportant functions can be derived from experiment.

V. UTILITY AND EXPERIMENTAL EVALUATIONOF THE OSMOTIC PRESSURE

We have repeatedly emphasized the importance and utility of the osmoticpressure � of foams and concentrated emulsions. Once known as a functionof �, it may be used to quantitatively link and predict a large numberof other important properties. Some of these are listed in this section. Inaddition, these considerations lead to a convenient method for evaluating�(�) experimentally (see Sec. V.D).

A. Motion of Continuous Phase Between Different Systemsin Contact

Let two concentrated dispersions with the same type of continuous phase[e.g., an aqueous foam and an oil-in-water emulsion, or two different o/wemulsions] be brought into contact, either directly or via a freely movablesemipermeable membrane. If the osmotic pressures are unequal (e.g., as aresult of differences in the volume fractions, mean drop size, interfacialtension, or combinations thereof), it is obvious that the (common) contin-uous phase will flow from the dispersion with the lower osmotic pressureinto that with the higher osmotic pressure until the two pressures are equal-ized. The final volumes and volume fractions of the two dispersions may bepredicted in a straightforward manner, once ~��ð�Þ is known. It is importantto point out that equality of the (mean) capillary pressures does not neces-sarily rule out flow nor does their inequality imply it.

B. Vapor Pressures of Continuous and Dispersed Phases

It can be shown (29) that the vapor pressure, pcv, of the continuous phase isreduced to below that of the bulk continuous phase, ð pcvÞ0, according to

pcv ¼ ð pcvÞ0 exp�� VV2

<T

� �

ð33Þ

where �VV2 is the partial molar volume of the solvent, < is the gas constant,and T is the absolute temperature.


Similarly, the vapor pressure of the dispersed phase, pd , in a concen-trated emulsion can be related to that of the bulk dispersed phase, ð pdv Þ0 by

pdv � ð pdv Þ0 exp2

R32

�VV1

<T

S

S0

� �

ð34Þ

where is the interfacial tension, R32 is the Sauter mean drop radius, �VV1 isthe molar volume of the dispersed liquid, and S/S0 is the relative increase insurface area at the volume fraction �. For �<�0, where S/S0¼ 1, werecover a variant of Kelvin’s equation; for �>�0, the increased vaporpressure is augmented further by the appearance of the factor S/S0 in theexponent, with S/S0 being related to �(�) through Eq. (31).

C. Gradient in / in Gravitational Field

So far, we have assumed that gravity is absent or negligible, so that thevolume fraction is uniform throughout the system. In gravity, however, asufficiently tall column will develop a significant gradient in � (24). Even ifeach individual drop is small enough to be essentially unaffected by the field(i.e., when the Bond number is very small), the combined buoyant forceof the underlying drops causes increasing drop deformation (and volumefraction) in the higher regions (Fig. 16). At the boundary between thedispersion and the bulk continuous phase, where z¼ 0, we have �¼�0,and the drops are purely spherical. At higher z, they increasingly deformuntil, as z ! 1, they acquire a purely polyhedral shape and �! 1. It is

Figure 16 Transition from spherical to polyhedral drops in vertical column. [From

Ref. 40. Copyright (1986) American Chemical Society.]


clear that, at any level, the combined buoyant force of all underlying dropsper unit area must equal the local osmotic pressure:

�ð�Þ ¼ ��g

Z z

0

� dz ð35Þ

or

~��ð�Þ ¼��gR32

Z z

0

�ðzÞ dz ¼

Z ~zz

0

�ð ~zzÞ d ~zz ð36Þ

where �� is the density difference between the phases, g is the accelerationdue to gravity, ~��ð�Þ is the reduced osmotic pressure

~��ð�Þ ¼�ð�Þ

=R32ð37Þ

and ~zz is the reduced height

~zz �R32z

a2cð38Þ

where ac¼ [/(��g)]1/2 is the capillary length.In all of the above, it is assumed that there is no gravitational segrega-

tion by drop size, that is, the drop size distribution does not vary with height.Thus, once ~��ð�Þ is known, �ð ~zzÞ can be evaluated from Eq. (36) in the

form

~zzð�Þ ¼

Z ~��

0

d ~��ð�Þ

�¼

Z �

�0

1

�

d ~��ð�Þ

d�

!

d� ð39Þ

As mentioned earlier, the only system for which ~��ð�Þ is known exactlyis the monodisperse 2D system [cf. Eq. (16)]. When Eq. (39) is applied to thiscase, we find

~zzð�Þ ¼1

ð�0�Þ1=2

1þ1� �01� �

� �1=2

ð2�� 1Þ

" #

� 2 ð40Þ

where �0¼ 0.9069. This result has been obtained also by Pacetti (64). Thevolume fraction profile is shown in Fig. 17.


D. Experimental Determination of ~&&ð/Þ for Real Systems

From the above, it is clear that ~��ð�Þ may be evaluated experimentally fromEq. (36) by determining the volume fraction as a function of height in anequilibrated (i.e., completely drained) dispersion column. This has beendone very carefully for a typical, well-characterized polydisperse emulsionof paraffin oil in water (24). The emulsion had a Sauter mean drop radius ofR32¼ 44.7 mm, an interfacial tension of 7.33mN/m, and a density differenceof 0.144 g/cm3. The experimental profile �ð ~zzÞ is given in Fig. 18 and may becompared with that in Fig. 17 for the monodisperse 2D system. It could benumerically fitted to the following equations, covering three differentranges of �:

‘‘Low’’-volume fraction (0.715<�<0.90 or 0< ~zz<0.5):

~zz ¼ 0:237�� 0:715

1� �

� �

ð41Þ

or

� ¼~zzþ 0:169

~zzþ 0:237ð42Þ

This leads to

~��ð ~zzÞ ¼ ~zz� 0:068 ln ð ~zzþ 0:237Þ � 0:098 ð43Þ

which, upon substitution for ~zz according to Eq. (41), leads to ~��ð�Þ.Equation (41) shows that �¼�0¼ 0.715 at ~zz¼ 0. This is one of our

Figure 17 Volume fraction versus reduced height for perfectly ordered 2D case.


reasons for concluding that typical polydisperse systems pack slightlyless tightly than ideally close-packed monodisperse systems, where�0¼ 0.7405.Intermediate-volume fraction (0.90<�<0.99 or 0.5< ~zz<4.0):

� ¼ 1:037 1� ð117:6 ~zzþ 4:0Þ�1=2� �

ð44Þ

and

~�� ¼0:00819�2

ð1� 0:9639�Þ2ð45Þ

High-volume fraction (0.99<�<1 or ~zz>4.0):

~zz � ~�� ¼ 0:58421� 1:892ð1� �Þ1=2� �2

ð1� �Þ1=2ð46Þ

which is the appropriate limiting solution for the polyhedral system.

Equations (42), (43), (45), and (46) describe the dependence of ~�� on �,as shown in Fig. 19. It may be compared with that for the monodisperse 2D

Figure 18 Experimental profile of volume fraction versus reduced height for typical

polydisperse emulsion. [From Ref. 24. Copyright (1987) American Chemical Society.]


and 3D systems in Figs. 5 and 15, respectively. Close examination showsthat the experimental osmotic pressure is consistently lower than those forthe idealized structures in Fig. 15.

Even though these relationships were derived for one particularemulsion, its size distribution was ‘‘typical,’’ so that we believe that theycan be applied with reasonable confidence in most practical situations.Nevertheless, more work remains to be done to elucidate the effect of thedetails of the size distribution. There is a particular need for the equivalentexpressions for the monodisperse system, which would serve as a benchmark.Bibette’s (65) novel way of preparing emulsions of low polydispersity(10% in radius) has opened up experimentation along these lines.Unfortunately, the technique appears to be capable only of generatingemulsions of extremely small drop size (R<1 mm), which complicatesmatters in several ways. First, estimates of the effective volume fractions[cf. Eq. (1)] become questionable, unless detailed quantitative information isavailable on the equilibrium film thickness as a function of the apparentvolume fraction (or capillary pressure). This is usually not the case, poten-tially leading to significant errors. Second, droplets of such small size areBrownian, which may lead to an entropic contribution to the osmotic

Figure 19 Reduced osmotic pressure as a function of volume fraction for typical

polydisperse emulsion. [From Ref. 24. Copyright (1987) American Chemical Society.]


pressure, in addition to the energetic contribution considered so far. Theseand other factors may be responsible for some of the differences between theabove results and those of Mason et al. (66), who measured �(�) for anoil-in-water ‘‘Bibette emulsion’’ of R¼ 0.48 mm. To cover the whole range of�, they used three different ways to generate the osmotic pressure: gravita-tional compaction, centrifugation, and dialysis of the emulsion against thecontinuous phase containing various levels of dextran, a polymer to whichthe dialysis membrane is impermeable. The osmotic pressure was found torise at an estimated effective � of (�0)e� 0.60 (rather than 0.715). This isclose to 0.64, the value for random close packing of uniform spheres. Up to�e¼ 0.80, the data could be fitted well to

~�� / �2ð�� 0:60Þ ð�<0:80Þ:

For �>0.80, the results of the two studies appear to be quite consis-tent, in spite of the disparity in the degree of polydispersity of the emulsionsemployed. The apparent discrepancy at the lower volume fractions may beentirely due to the large difference in mean drop size, for the reasons citedearlier.

E. Gravitational Syneresis or Creaming

In the absence of gravity (or with fluids of matched densities), a perfectlystable emulsion or foam with �> �0 will remain uniform and not ‘‘phaseseparate’’, (i.e., it will not exude a bottom layer of continuous phase). In agravitational (or centrifugal) field, such syneresis may occur, however, as aresult of compaction in the upper region (assuming that we are dealing witha foam or o/w emulsion; the continuous phase would separate at the top inw/o emulsions). In a consumer product, such behavior could be detrimental,as it might suggest instability, breakdown, and limited shelf life, even thoughsimple shaking would restore (temporary) uniformity. With the knowledgecontained in the previous subsection, it is possible to predict exactly whensuch syneresis will in fact occur (67). For a container of constant crosssection, the parameters of importance are the overall volume fraction, ��,and the reduced height of the sample, ~HH, defined by

~HH �HR32

a2c¼ HR32��

g

ð47Þ

where H is the actual height of the sample. It is clear that, for any ��, theremust be a critical reduced sample height, ~HHcr, above which syneresis will


occur and below which it will not. From a material balance and Eq. (36), itis readily shown that ~HHcr must obey the condition

~HHcr ¼~��ð ~HHcrÞ

�ð48Þ

Figure 20 shows how the resulting ~HH– �� diagram is bisected by ~HHcrð ��Þ.Reference 67 provides procedures for determining the height of the sepa-rated layer of continuous phase, if any, as well as the precise variation of �with height in the sample. The method may be extended to containers withvarying cross section (67). The following general conclusions may be drawn:(a) Everything else being equal, syneresis is less likely the higher the overallconcentration of the dispersed phase, ��; of course, when ��< �0, syneresiswill always occur; (b) For given �� (> �0), the tendency toward syneresis isless pronounced the smaller ~HH {i.e., for small drop size, high interfacialtension, small density difference, and small sample height [cf. Eq. (47)]};(c) For a foam or typical o/w emulsion, the tendency toward syneresis isreduced if the container is shaped with its widest part at the bottom. Thereverse is true for typical w/o emulsions.

F. Increase in Specific Surface Area with /

We have seen that the osmotic pressure is directly linked to the scaled specificsurface area, S/S0, as � increases from �0 through Eq. (31). For the mono-disperse 2D system, S/S0 is given by Eq. (13) and is plotted in Fig. 3.

Figure 20 Critical sample height for occurrence of syneresis as a function of

overall volume fraction.


To the extent that the real emulsion studied in Ref. 24 is representativeof typical polydisperse 3D systems, one can derive S/S0 from the expressionsfor ~��ð�Þ in Section V.D. The results (24) are as follows:

For 0.715<�<0.90,

S

S0¼ 1þ

1

3

0:084

��0:068

�lnð1� �Þ � 0:237

� �

ð49Þ

For 0.90<�<0.99,

S

S0¼

0:00283

1� 0:9639�þ 0:989 ð50Þ

For 0.99<�<1,

S

S0¼ 1:014þ 0:0686½1� 1:892ð1� �Þ1=2�3 ð51Þ

The combined results are shown in Fig. 21, where it is seen that the transi-tion from spheres to completely developed polyhedra is accompanied by anincrease in surface area of 8.3%. As mentioned earlier, for the monodispersecase, one predicts an increase in surface area of 9.7% on the basis of

Figure 21 Scaled specific surface area as a function of volume fraction for typical

polydisperse emulsion. [From Ref. 24. Copyright (1987) American Chemical

Society.]


Kelvin’s polyhedron as the ultimate drop shape, or 9.4% for the Weaire–Phelan structure. Polydispersity appears to give rise to an even somewhatsmaller overall change in surface area. Recent computer simulations ofvarious monodisperse and polydisperse structures by Kraynik et al. (68)confirm this result almost quantitatively.

G. Surface Area in Films Versus Total Surface Area

At any given volume fraction �, a fraction Sf/S of the total surface areaforms part of the films separating the droplets, and the remainder is still‘‘free’’ in the Plateau borders (Sf/S¼ 0 at �¼�0; Sf/S¼ 1 at �! 1). Thisparameter may play an important role in problems relating to the stabilityof, and mass transfer in, such systems. We have shown (29) that

Sf

S¼

S1=S0

S=S0

f ð�Þ

�2=3�

1:083

�2=3f ð�Þ

S=S0ð52Þ

where S/S0 is given by Fig. 21 and f(�) is the fraction of a confining wallthat is ‘‘contacted’’ by the flattened parts of the drops pushing against it,under the assumption that the wall is perfectly wetted by the continuousphase. This fraction, which varies from f¼ 0 at �0 to f¼ 1 at �¼ 1, can bemeasured experimentally (69) and was found empirically to be given by

f ð�Þ ¼ 1�3:20

�=ð1� �Þ þ 7:70ð Þ1=2

ð53Þ

for �0< �<0.975. (By solving for � at f¼ 0, we again obtain evidence that�0� 0.72 for real, polydisperse systems.) For �>0.975, we expect that f (�)is given, to a good approximation (40), by

f ð�Þ ¼ 1� 1:892ð1� �Þ1=2� �2

ð54Þ

Combining Eqs. (53) and (54) with Eq. (52) leads to the approximatedependence of Sf/S on � as shown in Fig. 22.

These are just some of the examples of where and how the osmoticpressure, or its related properties, can be used to define the overallequilibrium behavior of these complex fluids, even though their detailedmicroscopic structure may not be fully known. Other examples arepresented in Section VI, where we describe the only properties that areunique to foams as a result of the compressibility of their dispersed phase.


VI. FOAMS: INTERNAL PRESSURE, EQUATION OF STATE,AND COMPRESSIBILITY

Up to this point, we have emphasized the common structural and otherproperties of concentrated emulsions and foams. However, because oftheir gaseous dispersed phase, foams are compressible and, just as gasesthemselves, can be characterized by an equation of state that relates theirvolume, external pressure, and temperature.

A. Dry-Foam Limit (/¼ 1)

For a polydisperse dry foam, one can define an average internal pressure �ppithat is given by

�ppi ¼

PpiviPvi

¼

Ppivi

Vð55Þ

Figure 22 Fraction of total surface area contained in films as a function of volume

fraction for typical polydisperse emulsion. Solid curve at right is limiting solution for

fcc; the dashed curve connects it to the lower experimental region. [From Ref. 29.

Copyright (1988) American Chemical Society.]


where pi and vi are the pressure and volume of bubble i, respectively, and V isthe total foam volume. Derjaguin (70) has shown that

�ppi ¼ Pþ2S1

3Vð� ¼ 1Þ ð56Þ

where P is the external pressure and S1/V is the specific surface area of thefoam. Assuming ideality of the gas phase, this leads to the equation of state

Pþ2S1

3V

� �

V ¼ n<T ð� ¼ 1Þ ð57Þ

where n is the number of moles of gas in the foam. The same results werelater obtained by Ross (71).

Morrison and Ross (72) have indicated that although Eqs. (56) and(57) are undoubtedly correct for monodisperse foams, a rigorous proof oftheir validity for polydisperse systems was lacking. Such proof has since beenprovided by Hollinger (73), Crowley (74), and Crowley and Hall (75).

Derjaguin (70) further showed that the compression modulus K isgiven by

K � �VdP

dV¼

Pþ 2 �ppi

3¼ Pþ

4

9

S1

Vð� ¼ 1Þ ð58Þ

which compares to K¼P for a simple ideal gas.The specific surface area in Eqs. (56)–(58) may be replaced by

S1

V¼

S1

S0

S0

V¼

3

R32

S1

S0ð59Þ

where, as earlier, R32 is the Sauter mean bubble radius and S1/S0� 1.083 isthe increase in surface area associated with the transition from spherical topolyhedral bubbles at equal volume.

B. Foams with Finite Liquid Content (/ < 1)

We have shown (29) that, for this general case, Eqs. (56)–(58) are to bemodified as follows:

�ppi ¼ Pþ1� �

��þ

2S

3V1ð60Þ

Pþ1� �

��þ

2S

3V1

� �

�V ¼ n<T ð61Þ

K ¼1

�Pþ

1� �

3��þ ð1� �Þ2

d�

d�þ4

9

S

V1

� �

ð62Þ


where� is the osmotic pressure, V1 is the volume of the dispersed gas phase,and V is the total foam volume (V1¼�V). For �¼ 1, Eqs. (56)–(58) arerecovered.

Equations (60)–(62) may be written in the form

�ppi � P ¼

R32

1� �

�~��þ 2

S

S0

� �

ð63Þ

Pþ

R32

1� �

�~��þ 2

S

S0

� ��

�V ¼ n<T ð64Þ

K ¼1

�Pþ

R32

1� �

3�~��þ ð1� �Þ2

d ~��

d�þ4

3

S

S0

!" #

ð65Þ

where ~�� is the reduced osmotic pressure. The terms within the parenthesesdepend on � only and can be evaluated from the above-presented data.It may be shown (29) that the ‘‘osmotic’’ terms, although significant,provide only a rather small correction (<6%) to the dominant ‘‘Derjaguinterms’’ in S/S0. Of perhaps trivial but greater significance is the correctionfor the volume fraction outside the brackets of Eqs (61), (62), (64), and (65).

VII. MECHANICAL AND RHEOLOGICAL PROPERTIES

It has long been realized that the crowding of deformable drops and bubblesin concentrated emulsions and foams gives rise to interesting mechanicaland rheological properties, not shown by the separate constituent fluidphases. When subjected quasistatically to a small stress, these systemsrespond as purely elastic solids, characterized by a static elastic modulus,G. Under dynamic conditions, the modulus has a real, elastic component(the storage modulus, G0) and a complex, viscous component (the lossmodulus, G00). Once a critical or yield stress is exceeded, the systems flowand behave as viscoelastic fluids, whose effective viscosity decreases frominfinity (at the yield stress) with increasing shear rate. Thus, in rheologicalterms, they are plastic fluids with viscoelastic solid behavior below the yieldstress and viscoelastic fluid behavior above the yield stress.

A number of early experimental studies have provided qualitativeevidence for some or all of these behavioral aspects (e.g., Refs. 4 and76–82), but the techniques employed were usually crude and/or the systemswere poorly characterized, if at all. This makes it impossible to use theseearly experimental data to draw conclusions as to the quantitative relation-ships between the rheological properties, on the one hand, and important


system variables, such as volume fraction, interfacial tension, mean dropsize (and size distribution), fluid viscosities, shear rate, and so forth, on theother hand. In the last decade or so, interest in this area has intensified andmuch progress has been and is being made along several fronts: theoreticalmodeling, computer simulation, and careful experimentation. For otherrecent, although by now somewhat outdated, reviews, see Refs. 83–86.

A. Theoretical Modeling and Computer Simulation

In view of the exceedingly complex structure of 3D systems—even whenmonodisperse—initial efforts were confined almost exclusively to their 2Danalogs. Although unrealistic in some ways, these models provide importantkinematic insights and their behavior may be extrapolated, with cautionand limitations, to real systems. At first, for the sake of mathematical tract-ability, the complexity was reduced even further by considering perfectlyordered, monodisperse 2D systems. Gradually, the degree of complexity hasbeen increased by allowing disorder. It is only very recently that someintrepid investigators have begun to tackle the 3D problem in earnest.

1. Elastic and Yield Properties: Shear Modulus and Yield Stress

Two-Dimensional Systems. For the perfectly ordered case, theunstrained equilibrium structure has been discussed earlier. The (cylindrical)drops are arranged on a perfectly ordered hexagonal lattice, decorated at itsvertices with Plateau borders, whose wall curvatures are determined by thedrop size and volume fraction according to Eq. (11). The system can bethought to be confined between two parallel plates, with rows of drops beingforced to align with the plates. As one of the plates is now moved within itsown plane to induce shear, all drops respond by being deformed identically.In the process, the surface area increases. With the assumption of constantinterfacial tension, this results in a force (stress) versus deformation (strain)behavior that has been analyzed in detail, using straightforward geometricalarguments, by Princen (87) for any value of � � �0. The simplest dry-foamcase of �¼ 1 has been considered independently by Prud’homme (88).

The sequence of events in the dry-foam limit is illustrated in Fig. 23 fora single unit cell (i.e., the parallelogram formed by the centers of fouradjacent drops). As the cell is strained at constant volume, the angle betweenthe films must remain at 120�, which causes the central film to shorten untilits length shrinks to zero. At that point, four films meet in a line. Theresulting instability resolves itself by a rapid so-called T1 rearrangementor ‘‘neighbor switching.’’ In the process, new film is generated from thecenter to restore the original, unstrained configuration. A different, perhapsclearer, view of the system as it moves through such a cycle is shown in


Fig. 24. At any stage, the stress per unit cell is given by the horizontalcomponent of the tension of the originally vertical films; that is,

F ¼ 2 cos ð66Þ

where is the angle between these films and the horizontal shear direction.The resulting stress–strain curve per unit cell is given by curve 8 in Fig. 25,where ~FF is the dimensionless stress per unit cell:

~FF ¼F

2¼ cos ð67Þ

Figure 23 Shear deformation of unit cell of perfectly ordered 2D system in

dry-foam limit (�¼ 1); the transition from (c) to (d) is rapid and is often referred to

as a T1 rearrangement or neighbor switching. (From Ref. 87, with permission from

Academic Press.)


Figure 24 Alternative view of shear strain cycle. (From Ref. 87, with permission

from Academic Press.)

Figure 25 Shear stress per unit cell versus shear strain for perfectly ordered 2D

system at different volume fractions. (From Ref. 87, with permission from Academic

Press.)


Khan and Armstrong (45,89,90), using a slightly different analysis, arrivedat the following simple analytical result for curve 8:

~FF ¼�

ð�2 þ 4Þ1=2ð68Þ

where g is the imposed strain, which varies from zero to 2/ffiffiffi3

pat the point of

instability. The cycle then repeats itself.When �<1, the situation is considerably more complicated (Fig. 26).

As long as the two Plateau borders within the unit cell remain separated(Mode I), the stress per unit cell is unaffected. However, beyond a givenstrain, which depends on �, the Plateau borders merge to form a single,four-sided border. In this Mode II regime, the films no longer meet at 120�,and the stress–strain curve deviates from that for the dry-foam limit.

Figure 26 Increasing strain for systems with 0.9069<�<1. Between (a) and

(b), the system is in Mode I; between (b) and (c), the system is in Mode II. (From

Ref. 87, with permission from Academic Press.)


It passes through a (lower) maximum and ultimately reverses sign, eithercontinuously or via a T1 rearrangement (87). The resulting curves arecollected in Fig. 25. In each case, the maximum ~FFmax corresponds to thestatic yield stress per unit cell. It is plotted in Fig. 27 as a function of �,together with the corresponding yield strain. Realizing that there are 1/a

ffiffiffi3

p

unit cells per unit of length in the shear direction and that a may beexpressed in terms of the more practical drop radius R and volume fraction�, one finds for the stress(�)–strain(�) relationship

� ¼ 1:050

R�1=2 ~FFð�,�Þ ð69Þ

whereas the yield stress, �0, is given by

�0 ¼ 1:050

R�1=2 ~FFmaxð�Þ ð70Þ

Figure 27 Static yield stress per unit cell and yield strain as a function of volume

fraction for a perfectly ordered 2D system. (From Ref. 87, with permission from

Academic Press.)


where ~FFmaxð�Þ may be read from Fig. 27. It is expected to start deviatingfrom zero when adjacent layers of close-packed drops or bubbles can freelyslide past each other (i.e., at �¼�/4¼ 0.7854).

The small-strain, static shear modulus, G, is defined as

G �d�

d�

� �

�¼0

ð71Þ

and can be obtained from Eqs. (69) and (68):

G ¼ 1:050

R�1=2

d ~FF

d�

!

�¼0

¼ 0:525

R�1=2 ð�>�0Þ ð72Þ

The model predicts zero shear modulus for �<�0.Both the yield stress and the shear modulus scale with /R, but,

although the yield stress increases strongly with volume fraction, the shearmodulus is affected only very weakly through �1/2. In the dry limit of �¼ 1,both reach identical limiting values of

�0 ¼ G ¼ 0:525

Rð� ¼ 1Þ ð73Þ

The analysis may be extended to systems in which the film thickness, h,or the contact angle, �, between the films and the Plateau border walls arefinite (87). The effect of a finite film thickness is to increase the effectivevolume fraction [cf. Eq. (1)], which raises the yield stress and shear modulusin a predictable fashion. The effect of a finite contact angle on the shearmodulus is to simply reduce it by a factor of cos �. The effect on the yieldstress is more complex. In most but not all cases, the yield stress is increased.Furthermore, a finite contact angle can give rise to interesting new instabil-ity modes and to hysteretic behavior. The reader is referred to Ref. 87 forfurther details.

Subsequently, Khan and Armstrong (89,90) and Kraynik and Hansen(91) considered the effect of the orientation of the unit cell, relative to theshear direction, for the dry-foam case. They found that the shear modulusis unaffected, but that the yield stress is sensitive to the orientation. Inaddition, they considered planar extension as well as shear.

The sudden jump of the shear modulus from zero to a finite value at �0and its subsequent weak sensitivity to � for �>�0 are rather peculiar andappear to be associated with the perfect order of the model. The purecyclical character of the stress–strain curves is—by itself—a symptom of


‘‘perfection pathology.’’ Real systems do not exhibit these particular fea-tures, because they are invariably disordered, which causes T1 rearrange-ments to occur even at very small strains, as well as randomly throughoutthe system, rather than simultaneously at all vertices.

The shear modulus of polydisperse hexagonal systems of the typedepicted in Fig. 8b, is still given by Eq. (72) when R is replaced byRav¼ ð

PR2

i =nÞ1=2, a characteristic drop radius that is based on the average

drop area (46). However, as expected, the ‘‘elastic limit’’ (i.e., the stress andstrain where the first T1 rearrangement occurs) is reduced relative to that ofthe monodisperse case of the same volume fraction.

The elastic and yield properties of 2D systems with the most generaltype of disorder (cf. Fig. 7) have been simulated by Hutzler et al. (92) forboth dry and wet systems. Indeed, as the number of polydisperse drops inthe simulation is increased, the jumps in stress associated with individual orcooperative T1 rearrangements become less and less noticeable. Instead, thestress increases smoothly with increasing strain until it reaches a plateau thatmay be identified with the yield stress. The yield stress was found to increasesharply with increasing volume fraction, very much as in the monodispersecase. Furthermore, the shear modulus for the dry system (�¼ 1) was essen-tially identical to that for the monodisperse case, as given by Eq. (73) withRav, as defined earlier, replacing R. Its dependence on � was very differentfrom that in Eq. (72), however. When expressed in our terms, their resultsfor 1>�>0.88 could be fitted to

G

=Rav¼ 0:51� 21ð1� �Þ2 ð74Þ

Assuming that this relationship continues to hold for �<0.88 (where theirsimulations ran into difficulties because of the large number of T1 processesthe program had to deal with), the authors (92) concluded that G reacheszero at �¼�0� 0.84. As mentioned earlier, this ‘‘rigidity-loss transition’’can be identified as the random close packing of hard disks. The drop inG with decreasing � could further be correlated with the average number ofsides of the Plateau borders, which gradually increased from three close to�¼ 1 to about four at �¼ 0.84. Although these simulations involved arather small number of drops and leave some questions unanswered, theydo indicate a type of elastic behavior that—as we will see later—much moreclosely reflects that of real systems. Clearly, disorder plays a critical role.

Three-Dimensional Systems. The first expression for the shear

modulus of random dry foams (and emulsions) was derived by Derjaguin(93). It is based on the assumption that the foam is a collection of randomly


oriented films of constant tension 2 and negligible thickness and that eachfilm responds affinely to the applied shear strain, as would an imaginarysurface element in a continuum. Evaluating the contribution to the shearstress of a film of given orientation and averaging over all orientations thenleads to

G ¼4

15S1

Vð� � 1Þ ð75Þ

where S1/V is the surface area per unit volume. Because S1/V�

1.083S0/V¼ 3.25/R32, this may be written as

G �13

15

R32� 0:87

R32ð� � 1Þ ð76Þ

Much later, Stamenovic and Wilson (94) rediscovered Eq. (75), using similararguments but pointing out at the same time that it probably represents anoverestimate. Indeed, using 2D arguments, Princen and Kiss (95) concludedthat the affine motion of the individual films violates Plateau’s laws andleads to an overestimate of G by a factor of 2, at least in 2D. (Kraynik, in aprivate communication, pointed out an internal inconsistency in Ref. 95 andconcluded that G is overestimated by a factor of only 3/2). Furthermore,Derjaguin’s model does not allow for T1 rearrangements; it does not predicta yield stress nor does it have anything to say about the effect of � in ‘‘wet’’systems. On the other hand, the model correctly predicts that G scaleswith /R.

Stamenovic (96) analyzed the deformation of an idealized single foamvertex, where four Plateau borders meet and concluded that

G ¼1

6S1

V� 0:54

R32ð� � 1Þ ð77Þ

As pointed out by Reinelt and Kraynik (56), however, the idealized vertexdoes not adequately represent an equilibrium structure. Similar reserva-tions apply to the work of Budiansky and Kimmel (97), who consideredthe behavior of an isolated foam cell in the form of a regular pentagonaldodecahedron and obtained a shear modulus between the two above values.

Using Brakke’s Surface Evolver (49), Reinelt and co-workers(56,68,98–102) have explored in detail the elastic response of monodisperse,perfectly ordered structures, both ‘‘dry’’ and ‘‘wet,’’ to extensional and shearstrain. Structures considered included the rhombic dodecahedron, the regular(‘‘planar’’) tetrakaidecahedron, the Kelvin cell, and the Weaire–Phelan


structure. Some degree of disorder was introduced by considering bidisperseWeaire–Phelan systems (103), in which the relative volumes of the dodeca-hedra and tetrakaidecahedra were varied, as well as random, althoughmonodisperse, systems (68). As in the 2D case, the stress–strain behaviordepends on the cell orientation relative to the strain direction. Because of themultitude of edges and faces of each cell, a variety of T1 transitions mayoccur at increasing strain, leading to very complex behavior. Some of theirresults for the shear moduli of dry systems (�¼ 1) are listed in Table 1.

The ordered structures are all anisotropic, have cubic symmetry,and can be characterized by two shear moduli, G1 and G2. To simulateorientational disorder, the authors introduced an ‘‘effective isotropic shearmodulus,’’ Gav ¼ ð2=5ÞG1 þ ð3=5ÞG2, which is obtained by averaging over allorientations. The first three columns of Table 1 give the moduli in units ofV�1/3, where V is the cell volume; the last column is given in units of /R,where R¼ (3V/4�)1/3. The orientation-averaged results are surprisingly closeto the 2D prediction of G/R�1

¼ 0.525 [cf. Eq. (73)], Stamenovic’s predic-tion of G/R�1

¼ 0.54 [cf. Eq. (77)], and the extrapolated experimental resultof Princen and Kiss (95) for polydisperse emulsions, which indicated thatG/R�1

32 ¼ 0.509 (see Sect. VII.B.3). The small influence of polydispersity isalso suggested by the finding that Gav varies less than 0.5% when thevolume ratio of the two types of cells in bidisperse Weaire–Phelan struc-tures is varied between 0.039 and 2.392 (103).

Simulations of this type can pinpoint an ‘‘elastic limit’’ where the first(or subsequent) T1 transition(s) take(s) place. It depends extremely stronglyon orientation, as does the ‘‘dynamic yield stress’’ (i.e., the stress integratedover a complete strain cycle). The relevance to the yield stress of realdisordered systems is therefore quite limited (100). As in 2D simulations,simulations on more highly disordered systems will undoubtedly bringincreased insight.

Simulations on ‘‘wet’’ rhombic dodecahedra and Kelvin cells havebeen carried out by Kraynik and colleagues (68,102). The effective isotropicshear moduli were found to depend slightly on the volume fraction butdid not show the linear dependence on ��0 found experimentally for

Table 1 Shear Moduli of Dry Systems

G1/V�1/3 G2/V

�1/3 Gav/V�1/3 Gav/R

�1

Regular tetrakaidecahedron 0.5525 0.9696 0.8028 0.4980

Kelvin 0.5706 0.9646 0.8070 0.5006

Weaire–Phelan 0.8902 0.8538 0.8684 0.5387

Random (monodisperse) 0.78 0.08 0.48 0.05


disordered systems (95). Again, simulations on highly disordered wetsystems should improve our understanding.

Buzza and Cates (104) also addressed the question of whether disorderor the increased dimensionality from two to three dimensions is responsiblefor the observed experimental behavior of the shear modulus. In particular,they explored the lack of the sudden jump in G from zero to a finite value at�¼�0 that is predicted by the perfectly ordered 2D model. We have seenearlier that disorder appears to remove that abrupt jump in two dimensions(92). For drops on a simple cubic lattice, Buzza and Cates analyzed thedrop deformation in uniaxial strain close to �¼�0, first using the modelof ‘‘truncated spheres.’’ (For reasons given earlier, we believe this to be a verypoor model.) They showed that this model did not eliminate the discontinu-ous jump in G. An exact model, based on a theory by Morse and Witten(105) for weakly deformed drops, led to G / 1= lnð�� 0Þ, which gets rid ofthe discontinuity but still shows an unrealistically sharp rise at �¼�0 andis qualitatively very different from the experimentally observed lineardependence of G on �¼�0. Similar conclusions were reached by Lacasseand co-workers (51,106). A simulation of a disordered 3D model (106)indicated that the droplet coordination number increased from 6 at �0 to10 at �¼ 0.84, qualitatively similar to what is seen in disordered 2D systems(92). Combined with a suitable (anharmonic) interdroplet force potential,the results of the simulation were in close agreement with experimental shearmodulus and osmotic pressure data. Therefore, it appears again thatdisorder is responsible for many of the features of real systems.

2. Shear Viscosity

Compared to the quasistatic elastic and yield behaviors of concentratedemulsions and foams, the rate-dependent viscous properties are even morecomplex and relatively unexplored. Formally, the shear stress, �, may beexpressed as a function of the shear rate, _��, as

�ð _��Þ ¼ �0 þ �sð _��Þ ð78Þ

where �0 is the (elastic) yield stress and �sð _��Þ is the contribution from anyrate-dependent dissipative processes; or, in terms of the effective shearviscosity, me,

me ��ð _��Þ

_��¼�0_��þ�sð _��Þ

_��ð79Þ

The first term is, to a large extent, responsible for the shear-thinning behaviorof these systems. As is clear from the previous discussion, �0 is determinedprimarily by , R and �, whereas the size distribution may play a secondary


role. The dynamic stress, �s, is expected to depend on these and othervariables (e.g., the shear rate, the viscosities of the continuous and dispersedphases, and surface-rheological parameters). So far, the predictive qualityof theoretical and modeling efforts has been very restricted because of thecomplexity of the problem.

Buzza et al. (107) have presented a qualitative discussion of the variousdissipative mechanisms that may be involved in the small-strain linearresponse to oscillatory shear. These include viscous flow in the films,Plateau borders, and dispersed-phase droplets (in the case of emulsions),the intrinsic viscosity of the surfactant monolayers, and diffusion resistance.Marangoni-type and ‘‘marginal regeneration’’ mechanisms were consideredfor surfactant transport. They predict that the zero-shear viscosity is usuallydominated by the intrinsic dilatational viscosity of the surfactant mono-layers. As in most other studies, the discussion is limited to small-strainoscillations, and the rapid events associated with T1 processes in steadyshear are not considered, even though these may be extremely important.

It is now generally recognized that surfactants are indeed crucial, notonly in conferring (meta)stability to the emulsion or foam but also in con-trolling the rate-dependent rheology of the film surfaces and that of thesystem as a whole. Several early, spatially periodic 2D models neglectedthis aspect and made other simplifying assumptions. Khan and Armstrong(45,89,90) and Kraynik and Hansen (108) assumed that all of the continuousphase resides in the films (i.e., there were no Plateau borders) and that there isno exchange of fluid between the films. The film surfaces were assumed to becompletely mobile (no surfactant!). When such a system is strained globally,the uniform films respond with simple planar extension (or compression) atconstant volume. This mechanism predicts significant structural changes butleads to viscous terms in Eqs. (78) and (79) that are insignificant comparedwith the elastic terms up to extremely high shear rates that are unlikely to beencountered in practice. Experimentally, one finds a much more significantcontribution (see Sect. VII.B.3).

A more complete 2D analysis of simple shear is that of Li et al. (109).It solves the detailed hydrodynamics in the drops, films, and Plateau bordersfor the case of equal viscosities of the continuous and dispersed phases.Again, large structural changes are predicted. However, surfactants (andsurface tension gradients) are assumed to be absent, which severely limitsthe practical implications of the analysis. An interesting conclusion is that,under certain conditions, shear flow can stabilize concentrated emulsions,even in the total absence of surfactants.

An approach that is almost diametrically opposed to the earliermodels of Khan and Armstrong, and Kraynik and Hansen, was advancedby Schwartz and Princen (110). In this model, the films are negligibly thin,


so that all of the continuous phase is contained in the Plateau borders andthe surfactant turns the film surfaces immobile as a result of surface tensiongradients. Hydrodynamic interaction between the films and the Plateauborders is considered to be crucial. This model, believed to be more realisticfor common surfactant-stabilized emulsions and foams, draws on the workof Mysels et al. (111) on the dynamics of a planar, vertical soap film beingpulled out of, or pushed into, a bulk solution via an intervening Plateauborder. An important result of their analysis is commonly referred to asFrankel’s law, which relates the film thickness, 2h1, to the pulling velocity,U, and may be written in the form

h1

r¼ 0:643ð3Ca�Þ2=3 ð80Þ

where Ca*¼ mU/ (<<1) is the film-level capillary number, m and are theviscosity and surface tension of the liquid (the ‘‘continuous phase’’), and r isthe radius of curvature of the Plateau border where it meets the film and isgiven by capillary hydrostatics, r¼ (/2�g)1/2, where � is the density of theliquid and g is the gravitational acceleration.

Frankel’s law has its close analogs in a number of related problems(112–114) and has been verified experimentally (115,116) in the regimewhere the drawn-out film thickness, 2h1, is sufficiently large for disjoining-pressure effects to be negligible. Below some critical speed, the thickness ofthe drawn-out film equals the finite equilibrium thickness, 2heq, which isset by a balance of the disjoining pressure, �d (h), and the capillary pressure,/r, associated with the Plateau border. Thus, Frankel’s law and the follow-ing analysis apply only as long as 1>>Ca*2/3>> heq/r. It is expected tobreak down as the capillary number approaches zero. Disjoining pressureeffects may, in principle, be included (e.g., Ref. 117) but at the expense ofsimplicity and generality of the model.

The interesting hydrodynamics and the associated viscous energydissipation are confined to a transition region between the emerging, rigidlymoving film and the macroscopic Plateau border. The lubrication version ofthe Stokes equation may be used in this region, as the relative slope of theinterfaces remains small there.

It is reasonable to assume that the same basic process operates inmoving emulsions and foams. Lucassen (118) has pointed out that forsuch systems to be stable to deformations such as shear, the dilatationalmodulus of the thin films must bemuch greater than that of the surfaces in thePlateau border. However, this is equivalent to the assumption of inextens-ible film surfaces that underlies Frankel’s law. Therefore, it may well bethat, by implication, emulsions and foams that are stable to shear (and we


are interested in such systems only) have the appropriate surface rheologyfor Frankel’s law to apply. Of course, in emulsions and foams, each Plateauborder of radius r (set by drop size and volume fraction) is now shared bythree films. At any given moment, one or two of the films will be drawn outof the border while the other(s) is (are) pushed into it, at respective quasi-steady velocities U(t) that are dictated by the macroscopic motion of thesystem (Fig. 28). Using a perfectly ordered 2D system, Schwartz and Princen(110) considered a periodic uniaxial, extensional strain motion of smallfrequency and amplitude, so that inertial effects are negligible and compli-cations due to the merger of adjacent Plateau borders and associated rapidT1 processes are avoided. They proceeded by calculating the instantaneousrate of energy dissipation in the transition region of each of the three filmsassociated with a Plateau border and integrated the results over a completecycle. When the effective strain rate is related to the frequency of theimposed motion, the result can be expressed as an effective viscosity thatis given by*

me ¼ 5:3mCa�1=3ð81Þ

where the macroscopic capillary number Ca� ma _��/, a is the length of thehexagon that circumscribes a drop or bubble, and m is the viscosity of thecontinuous phase. Because of the small amplitude of the imposed motion,

Figure 28 Film being pulled out of a Plateau border with velocity U(t); all viscous

dissipation occurs in the transition region (II). (From Ref. 110, with permission from

Academic Press.)

*In the original article (110), the numerical coefficient was given as 6.7. This and a few other

minor numerical errors were pointed out by Reinelt and Kraynik (119, and private

communication).


the result does not depend on the volume fraction. It was further argued thatin the case of emulsions, the effect of the dispersed-phase viscosity, md, isrelatively insignificant. Reinelt and Kraynik (120) later estimated that this isa good approximation as long as

mdm<< Ca�1=3

ð82Þ

Apart from a change in the numerical coefficient, Eq. (81) is expected toapply also to a periodic, small-amplitude shearingmotion.However, in steadyshear, rapid film motions associated with the T1 processes, whose effect hasso far not been analyzed, periodically interrupt the above process. Further, asthe strain at the instability depends on the volume fraction (Fig. 27), theviscous term may become � dependent. Provided that the effect of the T1jumps may be neglected or the associated viscous contribution also scaleswith mCa�1/3, this model would then predict for the shear viscosity,

me ¼�0_��þ Cð�ÞmCa�1=3

¼�0_��þ C0ð�Þ

m2=31=3

R1=3_��1=3 ð83Þ

or, for the shear stress,

� ¼ �0 þ Cð�Þ

aCa2=3 ¼ �0 þ C0ð�Þ

m2=31=3

R1=3_��2=3 ð84Þ

where C(�) and C0(�) are of order unity and the yield stress �0 is given byEq. (70). Equations (83) and (84) describe a particular type of ‘‘Herschel–Bulkley’’ behavior, characterized in general by � ¼ �0 þ K _��n and me ¼�0= _�� þ K _��n�1. The special case of n¼ 1 is referred to as ‘‘Bingham plastic’’behavior. Occasionally, foams and concentrated emulsions are claimed tobehave as Bingham fluids. As we shall see, this is not so. (In fact, it isextremely unlikely that any fluid, when examined carefully, can be describedas such.)

Reinelt and Kraynik (119) improved on the above model by includingstructural changes that result from the fact that the film tensions deviatefrom the equilibrium value of 2 as they are being pulled out of or pushedinto the Plateau border. These changes are of order (Ca*)2/3, as alreadypointed out by Mysels et al. (111). As the values and signs of Ca* at anyinstant are different for the three films emanating from a Plateau border,their tensions are generally unequal and the angles between them deviatefrom 120�, and the Plateau border radius, r, is also affected. However, theserefinements do not alter the qualitative conclusion of the original model, asembodied in Eq. (81), for either planar-extensional or shear deformations.


Applying this approach to uniform dilatation of a foam, Reinelt andKraynik (119) also derived an expression for the dilatational viscosity,which again scales with mCa�1/3. Using a different surface-rheologicaldescription, Edwards and co-workers (121–123) arrived at alternativeexpressions for the dilatational viscosity of wet and dry foams.

In yet another extension, Reinelt and Kraynik (120) applied theapproach to steady shearing and planar-extensional flow of perfectly ordered2D systems for 0.9069<�<0.9466. This is the range of ‘‘very wet’’systems, for which the shear stress varies continuously with strain over acomplete strain cycle (cf. Fig. 25), so that rapid film events associated withT1 processes are avoided. They also investigated the effect of orientation,and structural effects due to changes in film tension were again included.As earlier, the effective viscosity was found to be proportional to mCa�1/3.Interestingly, the model indicates that the effective viscosity increases withincreasing volume fraction, which parallels practical experience.

Okuzuno and Kawasaki (124) simulated the shear rheology of dry,random 2D systems, using their ‘‘vertex model’’ in which the films areuncurved and do not generally meet at 120� angles. Although Plateau’scondition is therefore violated, the model offers the advantage of beingcomputationally more efficient than other, more realistic models. By solvingthe ‘‘equations of motion’’ for all the vertices, while taking account of T1rearrangements and using the energy-dissipation approach of Schwartz andPrincen (110), these authors tentatively concluded that the system behaveslike a Bingham plastic fluid. However, because the number of simulationswas quite limited, they did not rule out Herschel–Bulkley behavior withn 6¼ 1 (see above discussion). In a later study, the same investigators (125)observed violent flows like that of an avalanche in their simulations in thelarge strain regime at small shear rate. Similar avalanchelike flows areobserved in simulations by Jiang et al. (126).

This review is not exhaustive by any means. Other studies have beenand are being published regularly, as the topic continues to enjoy consider-able interest. It appears, however, that theoretical analyses and computersimulations can only go so far. There is a need for careful experimental workin order to establish the actual behavior of real systems. As has been the casein the past, further progress will be optimal when the two approaches gohand in hand.

B. Experimental Approaches and Results

The rheological parameters of primary scientific and practical concernare the static and dynamic shear modulus, the yield stress, and the shear-rate-dependent viscosity. The aim is to understand and predict how these


depend on the system parameters. In order to accomplish this with any hopeof success, there are two areas that need to be emphasized. First, the systemsstudied must be characterized as accurately as possible in terms of the volumefraction of the dispersed phase, the mean drop size and drop size distribu-tion, the interfacial tension, and the two bulk-phase viscosities. Second, therheological evaluation must be carried out as reliably as possible.

1. System Characterization

The bulk phases are generally Newtonian and their viscosities can bemeasured with great accuracy with any standard method available.

The nominal volume fraction of the dispersed phase can be obtainedvery accurately from the relative volumes (or weights) of the phases usedin the preparation of a highly concentrated emulsion (69). A series ofemulsions, differing only in volume fraction, may be conveniently preparedby dilution of a mother emulsion with varying known amounts of thecontinuous phase (69). Alternatively, if the phases differ greatly in volatility,the volume fraction may be obtained, albeit destructively, from the weightloss associated with evaporation of the more volatile phase, usually water(127). Another destructive method is to destroy the emulsion by high-speedcentrifugation in a precision glass tube, followed by accurate measurementof the relative heights of the separated liquid columns (24). To arrive atthe effective volume fraction, the nominal volume fraction may need to becorrected for a finite film thickness according to Eq. (1). Because allrheological parameters depend more or less strongly on the volume fraction,it is important that the vertical gradient in volume fraction due to gravity bekept to a minimum, if reliable rheological evaluations are to be expected.The gradient in volume fraction may be predicted quantitatively (67).Because the drop size and the density difference between the phases aregenerally much larger in foams than in emulsions, the gradient in � isusually much more pronounced in the former than in the latter. Therheologies of both types of system being governed by identical laws, it ispreferable—for this and many other reasons (see next paragraph)—to useemulsions, rather than foams, to learn about foam rheology.

The mean drop size and drop size distribution can be measured towithin a few percent accuracy with a number of techniques, such as the‘‘Coulter Counter’’ (69,95,128) and dynamic light scattering. The CoulterCounter is eminently suitable for oil-in-water emulsions but has a lowerpractical limit of about 1 mm. Various light-scattering techniques are equallysuitable for oil-in-water and water-in-oil (w/o) emulsions and afford a largerdynamic range. In either case, the concentrated emulsion must be dilutedwith the continuous phase to a level where coincidence counting or multiple


scattering, respectively, is avoided. One popular method that should perhapsbe avoided is optical microscopy, which is not only tedious but alsorelatively inaccurate when applied to polydisperse systems because ofdepth-of-focus limitations and wall effects. At any rate, a practical lowerlimit for accurate, quantitative optical microscopy is well in excess of 1 mm.Whatever method is used, it is desirable that complete size distributions bereported. At the very least, when only a mean drop size is reported, the typeof mean should be specified. Finally, it appears that size determinationsare much easier to obtain in emulsions than in foams. Moreover, althoughit is easy to prepare emulsions whose drop size distribution changesimperceptibly over a period of months, the bubble size distribution infoams changes very rapidly as a result of Ostwald ripening. It is, therefore,almost impossible to have accurate knowledge of the bubble size distribu-tion at the moment a rheological measurement is being made. These are yetadditional reasons for using emulsions in order to investigate foams.

The interfacial tension may be determined to within about 1%accuracy with the spinning drop method (129,130). It is an absolute andstatic method that requires only small samples and, in contrast to most othermethods, does not depend on the wettability of a probe, such as a ring orWilhelmy plate. The stabilizing surfactant is commonly used at concentra-tions in the bulk continuous phase that are far above the critical micelleconcentration (cmc). This ensures that the concentration remains above thecmc after adsorption onto the vastly extended interface has taken place,which is clearly needed to maintain emulsion stability. It is tempting,therefore, to assume that the interfacial tension in the finished emulsionequals that between the unemulsified bulk phases and that it remains con-stant when a ‘‘mother emulsion’’ is diluted with continuous phase in orderto create a series of emulsions in which only � is varied (69). This may be areasonable assumption when a pure surfactant is used, but there is evidencethat this may not be so when impure commercial surfactants are employed(95,128).

2. Rheological Evaluation

Most studies have used standard rheological techniques, such as rotationalviscometers of various types and geometries, such as concentric-cylinder,cone-and-plate, and parallel-plate rheometers, each of which may be opera-ted in various modes [constant stress, constant strain, steady shear, ordynamic (i.e., oscillatory) shear]. The relative advantages and/or limitationsof these and other techniques may be found in any standard textbook onpractical rheometry (e.g., Ref. 131). When applied to highly concentratedemulsions and foams—or suspensions in general, for that matter—these


techniques are fraught with many difficulties and pitfalls that are oftenoverlooked, leading to results of questionable validity. Some of thesedifficulties are the following.

Wall-induced Instability. Princen (69) has reported that otherwisevery stable oil-in-water emulsions showed extremely erratic behavior whensheared in a commercial concentric-cylinder viscometer with stainless-steelparts. The problem could be traced to the ‘‘coalescence’’ of the dispersed oildroplets with the steel walls and the formation of a thick oil layer.Apparently, the thin films of continuous phase separating the walls fromthe first layer of individual droplets were unstable and ruptured. Coating allrelevant parts with a thin film of silica, which assured adequate film stabilityand complete wetting of the steel by the continuous phase, solved theproblem (69). Later, an even more satisfactory solution consisted of replac-ing the steel inner and outer cylinders with glass parts, combined with otherimprovements in design (95,128,132). Some of the glass cylinders werehighly polished; others were roughened and equipped with vertical groovesto eliminate or reduce wall slip (see below). Wall-induced instability may ormay not be a problem, depending on the wall material, the emulsion (w/o oro/w), and surfactant type.

End and Edge Effects. In the analysis of raw data obtained withany type of rotational viscometer, it is assumed that the flow field isknown and simple. For example, in the conventional concentric-cylinderviscometer, it is assumed that the fluid moves in concentric cylindricallayers that extend unchanged from the precise top to the precise bottomof the inner cylinder. This is true only when the cylinders are infinitely long.For cylinders of finite length, complications at the top are usually minorand can often be neglected. In the lower region of the viscometer, however,the flow is seriously disturbed. In addition, the bottom of the innercylinder may contribute a substantial fraction of the total measuredtorque. This can lead to serious errors. Various suggestions have beenmade to deal with the problem (131), but their practical value is question-able. In addition to making other improvements, including the use of ahollow inner cylinder, Princen and Kiss (95,128,132) effectively isolatedthe bottom region by filling it with a layer of mercury. In that way, thesample of interest is strictly confined to the space between the cylinders.As long as its effective viscosity is much greater than that of mercury,flow between the cylinders is undisturbed and the torque on the bottomof the inner cylinder is negligible. The arrangement is shown schematicallyin Fig. 29.

In the cone-and-plate viscometer, there are similar, although perhapssomewhat less severe, problems associated with the outer edge (131).


Wall Slip. Along with wall-induced instability, the occurrence of slipbetween the sample and the viscometer walls is one of the most serious andprevalent, although often neglected, problems one encounters inassessing the rheology of dispersed systems, in general, and concentratedemulsions in particular. Because concentrated emulsions have a yield stress,wall slip—if present—can be readily demonstrated by painting a thin line ofdye on top of the sample in a wide-gap rotating-cylinder viscometer (69). Aslong as the yield stress is not exceeded at the inner cylinder wall, the sampleis not sheared at all but is seen to move around in the gap as an elasticallystrained solid! In this regime, shear is confined to the thin films of continu-ous phase separating the wall from the adjacent droplets. For a sufficientlysmooth wall, it is possible to estimate the thickness of these films from themeasured wall stress and angular velocity (69).

It is obvious that neglect of wall slip may lead to meaningless conclu-sions as to the system’s rheology. There are two different approaches todealing with this particular problem. First, one can try to eliminate slipby roughening the viscometer surfaces. Princen and Kiss (95) successfullyused roughened and grooved glass cylinders to determine the static shearmodulus of concentrated emulsions. This worked well in the low-stress,linear elastic regime, although, even here, some wall creep did occur(which could be readily corrected for). However, massive wall slip wasnoted to commence at shear stresses exceeding only about one-half of thebulk yield stress. Thus, even though the roughness was commensurate withthe drop size and served the intended purpose, the arrangement would havebeen inadequate for determining the yield stress and shear viscosity.Therefore, the question remains how rough a surface must be to eliminateslip up to the maximum shear stress considered. As an extreme case,large radial vanes have been recommended, at least for yield stress

Figure 29 Modified concentric-cylinder viscometer with glass outer cylinder,

hollow glass inner cylinder, and pool of mercury to confine the sample to the gap and

thus to minimize the end effect.


measurements (133). Although undoubtedly effective in preventing slip,the vanes do lead to some uncertainty in the strain field.

Many published rheological studies declare that wall slip was checkedfor and found to be absent. Unless solid evidence is provided, it behoovesthe reader to approach such assertions with a healthy dose of skepticism.

A second approach is to permit slip and to correct for it. This usuallyinvolves running the sample in two or more viscometer geometries (e.g., atdifferent gap widths) (131,134,135). Doubts have been expressed as to thevalidity of this approach (136). At any rate, the procedure is rather tediousand may not be very accurate. In an alternative method, Princen and Kiss(128), using their improved design with polished glass cylinders, establishedempirically that the torque versus angular velocity data for concentratedemulsions may be linearized over most of the all-slip/no-flow regime. Thestress at which the data deviated from this linear behavior was identified asthe yield stress. Under the further, reasonable assumption that the linearizedslip behavior persists above the yield stress, where flow commences, theangular velocity could be corrected for wall slip. Following standardrheological procedures for yield-stress fluids in a wide-gap concentric-cylinder viscometer, the dependence of the effective viscosity on shear ratecould then be determined.

It is clear from the above discussion that extreme care must be exer-cised in the characterization and rheological evaluation of concentratedemulsions. Few, if any, commercial viscometers are designed to give reliableresults for non-Newtonian fluids. Not only are modifications of the hard-ware often called for, but also the software of automated instruments isgenerally incapable of dealing with yield stress fluids, end effects, and wallslip. For example, to correct for end effects, it will not do to use a calibrationor ‘‘instrument factor’’ for any but Newtonian fluids. Unfortunately, thereare no shortcuts in this field!

3. Experimental Results

For reasons indicated earlier, accurate physical characterization andrheological evaluation of foams is extremely difficult. Indeed, althoughthere is much published material on foams that is qualitatively consistentwith what one would expect (and much that is not), we are not aware of anysuch studies that can stand close quantitative scrutiny. Therefore, we shallrestrict ourselves to what has been learned from highly concentratedemulsions, whose rheology is, in any case, expected to be identical to thatof foams in most respects. However, even in the emulsion area, the numberof carefully executed studies is severely limited. Admittedly not withoutsome prejudice, we shall concentrate on the systematic experimental work


by two groups that were active at different times at the Corporate ResearchLaboratory of Exxon Research and Engineering Co. [i.e., Princen andKiss (69,95,128) and Mason et al. (66,127,137,138)]. Both groups used oil-in-water emulsions, but whereas Princen and Kiss used ‘‘typical’’ polydis-perse emulsions with a mean radius of 5–10 mm, Mason and co-workersopted for submicron, monodisperse ‘‘Bibette emulsions.’’ The term ‘‘mono-disperse’’ is relative; there remained some polydispersity in drop radius ofabout 10% and the emulsions were structurally disordered on a macroscopicscale. The mean drop size in Princen’s emulsions was at least an orderof magnitude greater, which may account for some of the differencesin the results. Princen and Kiss used their customized concentric-cylinderviscometer exclusively, either in steady shear with wall slip (to give theyield stress and viscosity) or as a constant-strain device without wall slip(to give the static shear modulus). Mason and co-workers were moreeclectic in choosing their techniques (concentric-cylinder and cone-and-plate geometries in steady-shear and dynamic modes, as well as opticaltechniques).

Shear Modulus. Princen and Kiss (95) used a series of well-charac-terized, polydisperse oil-in-water emulsions of essentially identical Sautermean drop size, R32, and drop size distribution but varying dispersed-phasevolume fraction, �. Their modified Couette viscometer was purposelyequipped with ground and grooved glass cylinders to eliminate wall slip,*and the emulsion was strained by turning the outer cylinder over a small,precisely measured angle in the linear elastic regime. From the measuredstress at the inner cylinder, the static shear modulus, G, can be obtained ina straightforward manner. The results in Fig. 30 show that over the rangeconsidered (0.75<�<0.98), GR32/�

1/3 varies linearly with �, and we canwrite

G ¼ 1:77

R32�1=3ð�� 0:712Þ ¼ 1:77

R32�1=3ð�� 0Þ ð85Þ

where �0¼ 0.712 can be identified as the ‘‘rigidity-loss transition’’ for theparticular size distribution in these emulsions. This is surprisingly close tothat for ideal close packing of monodisperse spheres (�0¼ 0.7405) butclearly in excess of that for random close packing of monodisperse spheres(�0� 0.64). The exact value of �0 is expected to depend somewhat on thedetails of the drop size distribution.

*This fact was unfortunately misrepresented in Ref. 66.


In the ‘‘dry-foam’’ limit (�¼ 1), Eq. (85) reduces to

G�¼1 ¼ 0:509

R32ð86Þ

As indicated earlier, this is in close agreement with various theoreticalestimates.

It may be argued which mean drop size is most appropriate fordescribing the rheology of polydisperse systems. The selection of R32 isbased on limited evidence (69) and some other mean might ultimatelyturn out to be preferable. (See, however, the postscript in Sect. IX.)

A simple extension of the perfectly ordered 2D model to a 3D modelwould have suggested that G¼ 0 for �<�0¼ 0.74, with a sudden jump to analmost constant, finite value of G/ �1/3/R for �>0.74 [cf. Eq. (72)].As discussed earlier, it is now generally agreed that the absence of thediscontinuity and the essentially linear dependence on � above �0, foundexperimentally, is due to structural disorder.

Figure 30 Scaled static shear modulus, GR32/�1/3, versus � for typical

polydisperse emulsions. Solid points are experimental data; the solid line is drawn

according to Eq. (85). (From Ref. 95, with permission from Academic Press.)


Mason et al. (66) used small-amplitude, dynamic, oscillatory methods(both in cone-and-plate and concentric-cylinder geometries) to probe theviscoelastic properties [i.e., the storage (elastic) and loss (viscous) moduli,G 0 and G 00, respectively] as a function of frequency, !. No mention is madeof wall-induced instability or end and edge effects. Having roughened theviscometer walls, the authors claim that wall slip was nonexistent. At lowfrequencies, G 0 reached a plateau that may be equated with the static shearmodulus, G. Plots of the scaled modulus, GR/, versus the effective volumefraction, �e, for four emulsions of different drop size essentially overlapped,as expected. The drops were so small that significant corrections had to bemade to the nominal volume fractions to account for the finite (estimated)film thickness, h, according to Eq. (1). In the dry-foam limit (�e¼ 1), thescaled modulus approached a value of about 0.6, which is reasonably closeto Princen’s value of 0.51, but even for �<1, the data of the two groups areremarkably similar. For example, for �e¼ 0.85 and 0.75, Mason et al. showvalues for GR/ of about 0.30 and 0.10, respectively, whereas Eq. (85) yields0.23 and 0.061, respectively, for GR32/. The differences are roughlycommensurate with the scatter in Mason’s data. At any rate, the differencein polydispersity in the two sets of emulsions, or some experimental factor ineither study (end/edge effects?), may well explain these minor systematicdiscrepancies.

Overall, Mason et al. found that their data can be described by

G � 1:7

R� ð�� 0:64Þ ¼ 1:7

R� ð�� 0Þ ð87Þ

where �0� 0.64 is the value for random close packing of monodispersespheres. Except for the difference in �0, this is very similar to Eq. (85).

Because of the limited sensitivity of their viscometer and the increasedpotential effect of a gradient in � due to gravity, Princen et al. did notexplore the range of �<0.75 and reasonably assumed that the linearbehavior in Fig. 30 continues down to G¼ 0 at �¼ �0� 0.71. It is unclearwhat significance, if any, must be attached to the apparent difference in �0found in the two studies. Had it been possible to properly explore thatregime, Princen’s data might have shown some curvature for �<0.75 anda similar smooth decline in G toward zero at �0� 0.64. More likely, thedifference is real and simply attributable to the differences in polydispersityand associated random packing density. Another factor of potentialsignificance is the large difference in mean drop size. The drops in Mason’semulsions were submicron and, therefore, Brownian, which may contributean entropic (thermal) component to the modulus, as well as affect thepacking density.


Direct support for Eq. (85) has been reported by, among others, Taylor(139), Jager-Lezer et al. (140), Pal (141), and Coughlin et al. (142). Indirectsupport has been obtained by Langenfeld et al. (143), who compared thespecific surface areas of a number of water-in-oil emulsions as determinedby two independent methods: (a) from the measured shear modulus—whichyields R32 from Eq. (85), and thus the specific surface area from 3�/R32—and(b) from small-angle neutron scattering. The agreement was very satisfactory.

Yield Stress and Shear Viscosity. Using their modified concentric-cylinder viscometer—equipped in this case with polished glass inner andouter cylinders to allow unimpeded wall slip and a mercury pool to elimi-nate the lower end effect—Princen and Kiss (128) determined the yieldstresses, �0, and effective viscosities, með _��Þ, of a series of well-characterized,polydisperse oil-in-water emulsions. They empirically established that in allcases, the all-slip/no-flow regime at slow steady shear was characterized by alinear dependence of �1 on !/�1 (where �1 is the stress on the inner cylinderand ! is the angular velocity of the outer cylinder). The stress at which thedata deviated from this linearity was identified as the yield stress. At higherangular velocity, it was reasonably assumed that the same linear slipbehavior continued to operate, which permitted a straightforward slipcorrection. Using conventional rheometric analyses, the stress and viscositywere finally obtained as a function of shear rate.

The yield stress data could be expressed in the form

�0 ¼

R32�1=3Yð�Þ ð88Þ

The experimental values of Y(�) are shown in Fig. 31 and can be empiricallyfit to

Yð�Þ ¼ �0:080� 0:114 logð1� �Þ ð89Þ

Equation (89) should be used only within the range considered (i.e.,0.83<�<0.98).

Data by Pal (141) support Eqs. (88) and (89), once the volume fractionis corrected for a finite film thickness of 90 nm. Earlier data by Princen (69)are consistently somewhat higher, probably because of significant endeffects in the original, unmodified viscometer.

Figures 32 and 33 show the fully corrected plots of shear stress versusshear rate. Taking account of small differences in the measured interfacialtensions, all data could be accurately represented by

� ¼ �0 þ 32:0ð�� 0:73Þ

R32Ca1=2 ¼ �0 þ 32:0ð�� 0:73Þ

m _��R32

� �1=2

ð90Þ


where m is the viscosity of the continuous phase and Ca is the capillarynumber

Ca �mR32 _��

ð91Þ

which did not exceed a value of 10�4 in any of the experiments.For the effective viscosity, this leads to

me ��

_��¼�0_��þ 32:0ð�� 0:73ÞmCa�1=2

¼�0_��þ 32:0ð�� 0:73Þ

mR32 _��

� �1=2

ð92Þ

where �0 is given by Eqs. (88) and (89). Again, Eq. (92) should not be usedoutside the range considered. It is interesting to point out that, as with somany other properties, the viscous term tends to zero at �¼�0� 0.73.

It is encouraging that Eqs. (90) and (92) have the same form asEqs. (84) and (83), respectively, except for the exponent of the capillarynumber. Several reasons for this difference have been advanced (128),including the neglect of T1 rearrangements and disjoining pressure effects

Figure 31 Yield stress function Y(�)¼ �0R32/�1/3 versus � for typical polydisperse

emulsions. Solid points are experimental data; the curve is drawn according to

Eq. (89). (From Ref. 128, with permission from Academic Press.)


in the original model. At any rate, considering that this is the first and onlysystematic study of its kind, it is not yet clear how generally applicable Eqs.(90) and (92) will turn out to be. Although some other qualitative experi-mental support exists (144–146), there is a great need for additional, carefulstudies to explore this area further. It may be significant in this context thatLiu et al. (147), using diffusing-wave spectroscopy [a novel light-scattering

Figure 32 Fully corrected plots of shear stress vs shear rate for series of typical

polydisperse emulsions. Arrows indicate the yield stress, �0. Emulsions EM2–7 have

the same drop size (R32¼ 10.1 0.1 mm) but different volume fractions (�¼ 0.9706,

0.9615, 0.9474, 0.9231, 0.8889, and 0.8333, respectively). For EM8, R32¼ 5.73mmand �¼ 0.9474. (From Ref. 128, with permission from Academic Press.)


technique (148)] have found a contribution to the dynamic shear modulusthat is proportional to !1/2 (or Ca1/2) and increases roughly linearly withvolume fraction. Mason et al. (138) investigated the steady-shear behaviorof some monodisperse emulsions in the low-� range. They found thatthe viscous stress contribution varies as _��2=3 for �¼ 0.58 and as _��1=2

for �¼ 0.63. For �>0.65, no clear power-law behavior was observed.These authors claim that meaningful steady-shear measurements cannotbe made on emulsions of higher volume fractions because of the occurrenceof ‘‘inhomogeneous’’ strain rates. They presumably refer to the fact that, forexample, in a concentric-cylinder viscometer, only part of the emulsion(i.e., within a given radius) is being sheared, whereas the outer part is not.However, this situation, common to all yield-stress fluids, has been wellrecognized and analyzed in the rheology literature and can be handled ina quite straightforward manner (128).

Mason et al. (138) determined the yield stresses and yield strains of aseries of monodisperse emulsions, using either a cone-and-plate or double-wall Couette geometry in oscillatory mode. Wall-induced coalescence and

Figure 33 Plots of log(�� 0) versus shear rate for same emulsions as in Fig. 32.

In all cases, the slope is very close to 1/2. (From Ref. 128, with permission from

Academic Press.)


wall slip were claimed to be absent, but no mention is made of attempts toreduce end or edge effects. Estimated film thicknesses were used to arrive atthe effective volume fractions. Their data for the yield stress could be fit to

�0 ¼ 0:51

Rð�� 0:62Þ2 ð93Þ

and, for high �, are claimed to be ‘‘about an order of magnitude greater thanthose measured for polydisperse emulsions,’’ as given by Eqs. (88) and (89).This appears to be a misrepresentation. It is readily demonstrated that thetwo sets of data are, in fact, quite comparable. For example, for �¼ 0.85and �¼ 0.95, the values of the scaled yield stress, �0R/, are 0.027 and 0.055according to Eq. (93), and 0.013 and 0.067 according to Eq. (88). In fact, as�! 1, Mason et al. predicted that the scaled yield stress reaches a limitingvalue of 0.074, whereas extrapolation of Princen and Kiss’s data in Fig. 31suggest a value that is well in excess of 0.1 and perhaps as high as 0.15[the yield stress must remain finite in this limit and use of Eq. (89) isunwarranted in this regime]. Mason et al. further assert that, at high �,the yield strain of their monodisperse emulsions is also over an order ofmagnitude greater than that of the polydisperse emulsions of Princenand Kiss. This conclusion appears to be equally unfounded. In fact, therheological behavior of concentrated emulsions appears to be remarkablyunaffected by polydispersity.

We are not aware of any other systematic experimental studies thatmeet the criteria set out above and there remains a great need for additionalcareful work in this fascinating area.

VIII. ADDITIONAL AREAS OF INTEREST

Although this review covers many aspects of highly concentrated emulsionsand foams, it does not deal with a number of issues that are of considerableinterest. Foremost is the issue of emulsion and foam stability. A great deal ofinformation can be gleaned from recent books on foams and conventionalemulsions (17–22). Stability of highly concentrated emulsions is a rather moredelicate and specialized problem. The reader may consult a number of recentpublications that specifically deal with this subject (149–154).

One of the main driving forces for the recent upsurge in interest infoams—and one that has been responsible for the entrance of so manyphysicists into the field—has been their presumed usefulness in modelinggrain growth in metals. The coarsening of foam through gas diffusion(a special form of Ostwald ripening) is thought to follow similar laws.


This, among other things, inspired the first computer simulations of foamsby Weaire and co-workers and remains an active area of research (33).

As indicated, highly concentrated emulsions provide attractive startingmaterials for the synthesis of novel materials (e.g., polymers andmembranes).Ruckenstein has been particularly active in this area. In addition to thereferences cited earlier (6,12–16), the reader may wish to consult a recentcomprehensive review of this area (155).

IX. POSTSCRIPT

Except for some minor editing, this review is identical to Chapter 11 in Ref.156. Since then, further progress has been made. Two noteworthy examplesare the following.

In a recent tour de force, Kraynik, Reinelt, and van Swol (personalcommunication, 2002) have carried out numerous computer simulations onthe shear modulus of truly disordered, 3D, dry-foam systems, in which thedegree of polydispersity was varied by over two orders of magnitude. Theirwork comes to two important conclusions: (a) The surface volume or Sautermean radius R32 leads to a true numerical constant in Princen’s Eq. (86).The choice of other means, such as the simple average radius, appears tolead to a numerical factor that varies significantly with the degree of poly-dispersity. Moreover, Kraynik et al. have developed strong theoreticalgrounds for this conclusion. Although Princen’s initial choice of this parti-cular mean was based mostly on intuition, and on very limited experimentalevidence, that choice now appears to have been justified. (b) With R32 as themean, Kraynik et al. found that the appropriate value of the numericalfactor in Eq. (86) is 0.511, in remarkable agreement with Princen’s experi-mental value of 0.509! Because the error in both values is estimated to be afew percent, it can be stated with considerable confidence that the constantequals 0.51 0.01. Unfortunately, the case of ‘‘wet’’ foams (�<1) appearsto be computationally too intensive to be similarly simulated in the foresee-able future.

In a recent experimental study, Ponton et al. (157) measured theshear moduli of a series of laboratory-prepared, polydisperse water-in-oilemulsions of varying volume fraction. R32 was used as the characteristicmean radius. The results were in very close agreement with Eq. (85) in thatthe value of �0 was found to be 0.714, instead of Princen’s reported value of0.712. Again, we see remarkable agreement! Unfortunately, the authorsfailed to measure the interfacial tension in these emulsions, so that thevalue of the numerical factor in Eq. (85) could not be ascertained.


ACKNOWLEDGMENTS

Special thanks are due to A.M. Kraynik for the many stimulating discussionswe have had over the years, for keeping me informed on recent developments,and for kindly providing some of the unpublished results and illustrations.I have also benefited from illuminating discussions with P.-G. de Gennes andD. Weaire. My interest in these fascinating systems goes back to my years atUnilever Research, where E. D. Goddard and M. P. Aronson providedinvaluable and much appreciated support and collaboration.

LIST OF SYMBOLS

Latin Symbols

a Side of hexagon circumscribing compressed 2D drops in perfectorder

ao Side of hexagon circumscribing uncompressed (circular) 2D dropsin perfect order

ac Capillary length¼ [/(��g)]1/2

ci Mean curvature of surface between Plateau border and drop iCij Mean curvature of film between drops i and jCt Mean curvature of free surface of continuous phase at dispersion–

atmosphere boundaryCa Macroscopic capillary number¼ ma _��= or mR32 _��=Ca* Film-level capillary number¼ mU/e Number of edges of a polyhedral dropf Number of faces of a polyhedral dropf(�) Fraction of surface of confining wall ‘‘in contact’’ with dispersed

dropsF Stress per unit cellFmax Maximum or yield stress per unit cellg Acceleration due to gravityG Static shear modulusG0 Storage modulusG00 Loss modulush Film thicknessheq Equilibrium film thicknessh1 Half the film thickness pulled out of Plateau borderH Sample heightHcr Critical sample height for separation of continuous phaseK Compression moduluspb Pressure in Plateau border


pc Capillary pressurepi Pressure in drop ipcv Vapor pressure of continuous phase in dispersionð pcvÞ0 Vapor pressure of bulk continuous phasepdv Vapor pressure of dispersed phase in dispersionð pdv Þ0 Vapor pressure of bulk dispersed phaseP External pressurer Radius of Plateau border surfaces in 2D close-packed dropsR Radius of spherical or circular dropRav Average drop radiusR32 Surface volume or Sauter mean drop radius< Gas constantS Surface area of compressed dropsS0 Surface area of uncompressed (spherical or circular) dropsSf Surface area contained in filmsT Absolute temperatureU Film velocityv Number of vertices of a polyhedral dropV Dispersion volumeV1 Volume of the dispersed phaseV2 Volume of the continuous phase in the dispersion�VV1, �VV2 Partial molar volume of phase 1 and 2, respectivelyY(�) Yield-stress functionz Vertical height in dispersion column

Greek Symbols

� Strain_�� Rate of strain�� Density difference� Contact angle at film–Plateau border junctionm Viscosity of continuous phasemd Viscosity of dispersed phaseme Effective viscosity of dispersion� Osmotic pressure�d Disjoining pressure� Density Surface or interfacial tension� Stress�0 Yield stress�s Stress due to dissipative processes� Volume fraction of dispersed phase in emulsion or foam


�0 Volume fraction of close-packed spherical drops�e Effective volume fraction, after correction for finite film thickness Angle between films and shear direction! Frequency or angular velocity

REFERENCES

1. F. J. Almgren and J. E. Taylor, Sci. Am. 235, 82 (1976).

2. C. S. Smith, Metall. Rev. 9, 1 (1964).

3. E. B. Matzke, Am. J. Bot. 33, 58, 130 (1946).

4. K. J. Lissant, J. Colloid Interf. Sci. 22, 462 (1966).

5. K. J. Lissant, J. Soc. Cosmet. Chem. 21, 141 (1970).

6. K. J. Lissant and K. G. Mayhan, J. Colloid Interf. Sci. 42, 201 (1973).

7. A. Beerbower, J. Nixon, and T. J. Wallace, J. Aircraft 5, 367 (1968).

8. A. Beerbower, J. Nixon, W. Philippoff, and T. J. Wallace, SAE Trans. Sec. 2 76,

1446 (1968).

9. J. Nixon, A. Beerbower, and T. J. Wallace, Mech. Eng. 90, 26 (1968).

10. J .Nixon, and A. Beerbower, Preprints, Div. Petrol. Chem. Am. Chem. Soc. 14,

49 (1969).

11. K. J. Lissant, U.S. patent No. 3,892,881.

12. K. J. Lissant, B. W. Peace, S. H. Wu, and K. G. Mayhan, J. Colloid Interf. Sci.

47, 416 (1974).

13. J. M. Williams, Langmuir 4, 44 (1988).

14. E. Ruckenstein and K. Kim, J. Appl. Polym. Sci. 36, 907 (1988).

15. E. Ruckenstein, Colloid Polym. Sci. 267, 792 (1989).

16. E. Ruckenstein and J. S. Park, Chem. Mater. 1, 343 (1989).

17. J. Sjoblom (ed.), Emulsions and Emulsion Stability, Surfactant Science Series

Vol. 61, Marcel Dekker, New York; 1996.

18. R. K. Prud’homme and S. A. Khan (eds.), Foams: Theory, Measurements, and

Applications, Surfactant Science Series Vol. 57, Marcel Dekker, New York,

1996.

19. L. L. Schramm (ed.), Foams: Fundamentals and Applications in the Petroleum

Industry, Advances in Chemistry Series No. 242, American Chemical Society,

Washington, DC, 1994.

20. D. Exerowa and P. M. Kruglyakov (eds.), Foam and Foam Films: Theory,

Experiment, Application. Studies in Interface Science Vol. 5, Elsevier,

Amsterdam, 1998.

21. D. Weaire and S. Hutzler, The Physics of Foams, Clarendon Press, Oxford,

1999.

22. J. F. Sadoc and N. Rivier (eds.) Foams and Emulsions, Kluwer Academic,

Dordrecht, 1999.

23. H. M. Princen, M. P. Aronson, and J. C. Moser, J. Colloid Interf. Sci. 75,

246, (1980).

24. H. M. Princen and A. D. Kiss, Langmuir 3, 36 (1987).


25. D. M. A. Buzza and M. E. Cates, Langmuir 9, 2264 (1993).

26. M. P. Aronson and H. M. Princen, Nature 286, 370 (1980).

27. M. P. Aronson and H. M. Princen, Colloids Surfaces 4, 173 (1982).

28. H. M. Princen, Colloids Surfaces 9, 47 (1984).

29. H. M. Princen, Langmuir 4, 164 (1988).

30. A. V. Neimark and M. Vignes-Adler, Phys. Rev. E 51, 788 (1995).

31. F. Bolton and D. Weaire, Phil. Mag. B 63, 795 (1991).

32. V. V. Krotov and A. I. Rusanov, Mendeleev Commun. 177 (1998).

33. J. A. Glazier and D. Weaire, J. Phys.: Condens. Matter 4, 1867 (1992).

34. J. A. Glazier and J. Stavans, Phys. Rev. A 40, 7398 (1989).

35. J. A. Glazier, S. P. Gross, and J Stavans, Phys. Rev. A 36, 306 (1987).

36. J. Stavans and J. A. Glazier, Phys. Rev. Lett. 62, 1318 (1989).

37. J. Stavans, Phys. Rev. A 42, 5049 (1990).

38. J. Lucassen, S. Akamatsu, and F. Rondelez, J. Colloid Interf. Sci. 144,

434 (1991).

39. H. M. Princen, J. Colloid Interf. Sci. 71, 55 (1979).

40. H. M. Princen, Langmuir 2, 519 (1986).

41. P. M. Kruglyakov, D. R. Exerowa, and K. I. Khrystov, Langmuir 7; 1846

(1991).

42. K. Khrystov, P. Kruglyakov, and D. Exerowa, Colloid Polym. Sci. 257, 506

(1979).

43. K. Khrystov, D. Exerowa, and P. M. Kruglyakov, Colloid J. (Engl. Transl.) 50,

765 (1988).

44. D. Weaire and J. P. Kermode, Phil. Mag. B 50, 379 (1984).

45. S. A. Khan and R. C. Armstrong, J Non-Newtonian Fluid Mech. 25, 61 (1987).

46. A. M. Kraynik, D. A. Reinelt, and H. M. Princen, J. Rheol 35, 1235 (1991).

47. F. Bolton and D. Weaire, Phil. Mag. B 65, 473 (1992).

48. S. Hutzler and D. Weaire, J. Phys: Condens. Matter. 7, L657 (1995).

49. K. A. Brakke, Exp. Math. 1, 141 (1992).

50. A. M. Kraynik and D. A. Reinelt (private communication; to be published).

51. M.-D. Lacasse, G. S. Grest, and D. Levine, Phys. Rev. E 54, 5436 (1996).

52. W. Thomson (Lord Kelvin), Phil. Mag. 24, 503 (1887).

53. W. Thomson (Lord Kelvin), Acta Math. 11, 121 (1887–1888).

54. W. Thomson (Lord Kelvin), Mathematical and Physical Papers, Cambridge

University Press, London, 1911, Vol. V, p 297.

55. H. M. Princen and P. Levinson. J. Colloid Interf. Sci. 120, 172 (1987).

56. D. A. Reinelt and A. M. Kraynik, J. Colloid Interf. Sci. 159, 460 (1993).

57. D. Weaire and R. Phelan, Phil. Mag. Lett. 69, 107 (1994).

58. S. Ross and H. F. Prest, Colloids Surfaces 21, 179 (1986).

59. D. S. Bohlen, H. T. Davis and L. E. Scriven, Langmuir 8, 892 (1992).

60. H. W. Schwartz, Rec. Trav. Chim. 84, 771 (1964).

61. H. Aref, and T. Herdtle, in Topological Fluid Mechanics, (H. Moffat and

A. Tsinober, eds.), Cambridge University Press, New York, 1990, p. 745.

62. T. Herdtle, Ph.D. thesis, University of California, San Diego, 1991.

63. C. Monnereau and M. Vignes-Adler, Phys. Rev. Lett. 80, 5228 (1998).


64. S. D. Pacetti, Masters thesis, University of Houston, 1985.

65. J. Bibette, J. Colloid Interf. Sci. 147, 474 (1991).

66. T. G. Mason, M.-D. Lacasse, G. S. Grest, D. Levine, J. Bibette, and

D. A. Weitz, Phys. Rev. E 56, 3150 (1997).


68. A. M. Kraynik, M. K. Neilsen, D. A. Reinelt, and W. E. Warren, in Foams and

Emulsions. (J. F. Sadoc and N. Rivier, eds.), Kluwer Academic, Dordrecht,

1999, pp. 259–286.


70. B. V. Derjaguin, Kolloid Z. 64, 1 (1933).

71. S. Ross, Ind. Eng. Chem. 61, 48 (1969).

72. I. D. Morrison and S. Ross, J. Colloid Interf. Sci. 95, 97 (1983).

73. H. B. Hollinger, J. Colloid Interf. Sci. 143, 278 (1991).

74. T. L. Crowley, Langmuir 7, 430 (1991).

75. T. L. Crowley and D. G. Hall, Langmuir 9, 101 (1993).

76. M. Blackman, Trans. Faraday Soc. 44, 205 (1948).

77. J. O. Sibree, Trans. Faraday Soc. 30, 325 (1934).

78. A. David and S. S. Marsden, in 44th Annual Meeting, Society of Petroleum

Engineers, 1969, paper 2544.

79. V. Sanghani and C. U. Ikoku, Trans. ASME 105, 362 (1983).

80. B. J. Mitchell, Oil Gas J. 96 (1971).

81. S. H. Raza and S. S. Marsden, Soc. Petrol. Eng. J. 7, 359 (1967).

82. R. J. Mannheimer, J. Colloid Interf. Sci. 40, 370 (1972).

83. J. P. Heller and M. S. Kuntamukkula, Ind. Eng. Chem. 26, 318 (1987).

84. J. H. Aubert, A. M. Kraynik and P. B. Rand, Sci. Am. 254, 58 (1986).

85. A. M. Kraynik, Annu. Rev. Fluid Mech. 20, 325 (1988).

86. D. Weaire and M. A. Fortes, Adv. Phys. 43, 685 (1994).


88. R. K. Prud’homme, in Annual Meeting of the Society of Rheology, 1981.

89. S. A. Khan, Ph.D. thesis, Massachusetts Institute of Technology, 1985.

90. S. A. Khan and R. C. Armstrong, J. Non-Newtonian Fluid Mech. 22, 1 (1986).

91. A. M. Kraynik and M. G. Hansen, J. Rheol. 30, 409 (1986).

92. S. Hutzler, D. Weaire, and F. Bolton, Phil. Mag. B 71, 277 (1995).

93. B. Derjaguin, Kolloid Z. 64, 1 (1933).

94. D. Stamenovic and T. A. Wilson, J. Appl. Mech. 51, 229 (1984).

95. H. M. Princen and A. D. Kiss, J. Colloid Interf. Sci. 112, 427 (1986).

96. D. Stamenovic, J. Colloid Interf. Sci. 145, 255 (1991).

97. B. Budiansky and E. Kimmel, J. Appl. Mech. 58, 289 (1991).

98. D. A. Reinelt, J. Rheol. 37, 1117 (1993).

99. A. M. Kraynik and D. A. Reinelt, Forma 11, 255 (1996).

100. D. A. Reinelt and A. M. Kraynik, J. Fluid Mech. 311, 327 (1996).

101. A. M. Kraynik and D. A. Reinelt, J. Colloid Interf. Sci. 181, 511 (1996).

102. A. M. Kraynik and D. A. Reinelt, Proceedings of the XIIth International

Congress on Rheology, 1996, p. 625.

103. A. M. Kraynik and D. A. Reinelt, Chem. Eng. Commun. 148/150, 409 (1996).


104. D. M. A. Buzza and M. E. Cates, Langmuir 10, 4502 (1994).

105. D. C. Morse and T. A. Witten, Europhys. Lett. 22, 549 (1993).

106. M.-D. Lacasse, G. S. Grest, D Levine, T. G. Mason, and D. A. Weitz, Phys.

Rev. Lett. 76, 3448 (1996).

107. D.M. A. Buzza, C.-Y. D. Lu, andM. E. Cates, J. Phys. II (France) 5, 37 (1995).

108. A. M. Kraynik and M. G. Hansen, J. Rheol. 31, 175 (1987).

109. X. Li, H. Zhu and C. Pozrikidis, J. Fluid. Mech. 286, 379 (1995).

110. L. W. Schwartz and H. M. Princen, J. Colloid Interf. Sci. 118, 201 (1987).

111. K. J. Mysels, K. Shinoda, and S. Frankel, Soap Films: Studies of Their

Thinning and a Bibliography, Pergamon Press, New York, 1959.

112. L. Landau and B. Levich, Acta Physicochim. URSS 17, 42 (1942).

113. F. P. Bretherton, J. Fluid Mech. 10, 166 (1961).

114. L. W. Schwartz, H. M. Princen, and A. D. Kiss, J. Fluid Mech. 172, 259 (1986).

115. K. J. Mysels and M. C. Cox, J. Colloid Interf. Sci. 17, 136 (1962).

116. J. Lyklema, P. C. Scholten, and K. J. Mysels, J. Phys. Chem. 69, 116 (1965).

117. G. F. Teletzke, H. T. Davis, and L. E. Scriven, Rev. Phys. Appl. 23, 989 (1988).

118. J. Lucassen, in Anionic Surfactants: Physical Chemistry of Surfactant Action.

(E. H. Lucassen-Reijnders, ed), Surfactant Science Series Vol. 11, Marcel

Dekker, New York, 1981, p. 217.

119. D. A. Reinelt and A. M. Kraynik, J. Colloid Interf. Sci. 132, 491 (1989).

120. D. A. Reinelt and A. M. Kraynik, J. Fluid Mech. 215, 431 (1990).

121. D. A. Edwards, H. Brenner, and D. T. Wasan, J. Colloid Interf. Sci. 130, 266

(1989).

122. D. A. Edwards and D. T. Wasan, J. Colloid Interf. Sci. 139, 479 (1990).

123. D. A. Edwards and D. T. Wasan, in Foams: Theory, Measurements, and

Applications. (R. K. Prud’homme and S. A. Khan, eds), Surfactant Science

Series Vol. 57, Marcel Dekker, New York, 1996, p189.

124. T. Okuzuno and K. Kawasaki, J. Rheol. 37, 571 (1993).

125. T. Okuzono and K. Kawasaki, Phys. Rev. E 51, 1246 (1995).

126. Y. Jiang, P. J. Swart, A. Saxena, M. Asipauskas, and J. A. Glazier, Phys. Rev.

E 59, 5819 (1999).

127. T. G. Mason, J Bibette, and D. A. Weitz, Phys. Rev. Lett. 75, 2051 (1995).

128. H. M. Princen and A. D. Kiss, J. Colloid Interf. Sci. 128, 176 (1989).

129. H. M. Princen, I. Y. Z. Zia, and S. G. Mason, J. Colloid Interf. Sci. 23,

99 (1967).

130. J. L. Cayias, R. S. Schechter, and W. H. Wade. in Adsorption at Interfaces

(K. L. Mittal, ed), ACS Symposium Series No. 8, American Chemical Society,

Washington, DC, 1975, p. 234.

131. R. W. Whorlow, Rheological Techniques, Ellis Horwood, Chichester, 1980.

132. H. M. Princen, J. Rheol. 30, 271 (1986).

133. P. V. Liddell and D. V. Boger, J. Non-Newtonian Fluids 63, 235 (1996).

134. A. S. Yoshimura and R. K. Prud’homme, Soc. Petrol Eng. 735 (1988).

135. A. S. Yoshimura and R. K. Prud’homme, J. Rheol. 32, 53 (1988).

136. P. Brunn, M. Muller and S. Bschorer, Rheol. Acta 35, 242 (1996).

137. T. G. Mason and D. A. Weitz, Phys. Rev. Lett. 74, 1250 (1995).


138. T. G. Mason, J. Bibette, and D. A. Weitz, J. Colloid Interf. Sci. 179, 439 (1996).

139. P. Taylor, Colloid Polym. Sci. 274, 1061 (1996).

140. N. Jager-Lezer, J.-F. Tranchant, V. Alard, C. Vu, P. C. Tchoreloff, and

J.-L. Grossiord, Rheol. Acta 37, 129 (1998).

141. R. Pal, Colloid Polym. Sci. 277, 583 (1999).

142. M. F. Coughlin, E. P. Ingenito, and D Stamenovic, J. Colloid Interf. Sci. 181,

661 (1996).

143. A. Langenfeld, F. Lequeux, M.-J. Stebe, and V. Schmitt, Langmuir 14,

6030 (1998).

144. F. van Dieren, in Theoretical and Applied Rheology. (P. Moldenaers and

R. Keunings, eds.), Proceedings of the XIth International Congress on Rheology

Elsevier, Amsterdam 1992, p. 690.

145. Y. Otsubo and R. K. Prud’homme, Soc. Rheol. (Japan) 20, 125 (1992).

146. Y. Otsubo and R. K. Prud’homme, Rheol. Acta 33, 303 (1994).

147. A. J. Liu, S. Ramaswamy, T. G. Mason, H. Gang, and D. A. Weitz, Phys. Rev.

Lett. 76, 3017 (1996).

148. D. J. Pine, D. A. Weitz, P. M. Chaikin, and E. Herbolzheimer, Phys. Rev. Lett.

60, 1134 (1988).

149. E. Ruckenstein, G. Ebert, and G. Platz, J. Colloid Interf. Sci. 133, 432 (1989).

150. H. H. Chen and E. Ruckenstein, J. Colloid Interf. Sci. 138, 473 (1990).

151. H. H. Chen and E. Ruckenstein, J. Colloid Interf. Sci. 145, 260 (1991).

152. M. P. Aronson, K. Ananthapadmanabhan, M. F. Petko, and D. J. Palatini,

Colloids Surfaces A 85, 199 (1994).

153. M. P. Aronson and M. F. Petko, J. Colloid Interface Sci. 159, 134 (1993).

154. J. Bibette, D. C. Morse, T. A. Witten and D. A. Weitz, Phys. Rev. Lett. 69,

2439 (1992).

155. E. Ruckenstein, Adv. Polym. Sci. 127, 1 (1997).

156. H. M. Princen, in Encyclopedic Handbook of Emulsion Technology. (J.

Sjoblom, ed.), Marcel Dekker, New York, 2000, p. 243.

157. A. Ponton, P. Clement, and J. L. Grossiord, J. Rheol. 45, 521 (2001).


12Beverage Emulsions

Chee-Teck TanConsultant, Middletown, New Jersey, U.S.A.

I. INTRODUCTION

Since the late 1990s, the beverage market in the United States had gonebeyond providing beverages for refreshment to health and wellness claims.Beverages are also found to be the most successful and popular means ofdelivering functional benefits (1). These new beverages are flavored tea,flavored water, juice drinks, dairy-based juice drinks, and fortified beverages.In the United States, by the flavor category, the market statistic shows colaranks first, and lemon-lime and orange are second and third, respectively.However, worldwide, orange flavor is the favorite of consumers (2). Thesecitrus flavor products are based primarily on the essential oils from the peelof the fruits (e.g., orange oils, lemon oils, etc.). Like all essential oils, theyare not water miscible. They cannot be mixed directly with a sugar solutionto make a beverage. There are two methods of incorporating these flavorsinto soft drinks. The first method involves separating out the water-solublefraction of the flavor from the essential oils by extraction and distillation.The second method involves converting the oil into a water-dispersibleemulsion (i.e., a beverage emulsion).

Beverage emulsions are a unique class of emulsions. They are differentfrom other food emulsions in that they are to be consumed in a highlydiluted form, rather than in their original concentrate form. They arefirst prepared as an emulsion concentrate, which is later diluted in a sugar


solution to produce the finished beverage—the soft drink. The soft drinkprepared can be either a noncarbonated still drink or a carbonated drink.In soft drinks, the beverage flavor emulsion provides flavor, color, andcloudy appearance for the beverage, or just simply the cloudiness whenthe beverage cloud emulsion is used. These characteristics of beverageemulsions are what make them a unique class of emulsions. In the prepara-tion of the finished beverages, the beverage emulsion concentrate is dilutedin a sugar solution to produce the finished soft drinks; it is usuallydiluted several hundred times. The emulsions in both the concentrate anddiluted forms must have a high degree of stability. They must be stablefor at least 6 months, as required by the beverage industry. Other examplesof emulsions consumed in a diluted form are coffee whitener and dairycream for coffee. However, these emulsions are consumed almost immedi-ately after dilution; the stability of the diluted form is required only for ashort time. Therefore, dairy cream and coffee whiteners do not have thesame stability problem as encountered by beverage emulsions for softdrinks. In this chapter, we will deal mainly with beverage emulsions forsoft drinks.

II. DEFINITION

Beverage emulsions can be divided into two categories: beverage flavoremulsions and beverage cloud emulsions. Beverage flavor emulsions providethe beverage with flavor, cloudiness, and color as in certain formulas.Beverage cloud emulsions provide only cloudiness with no flavor. Bothbeverage emulsions are composed of an oil phase and a water phase, andthey are classified as oil-in-water (o/w) emulsions. The oil phase consistsof flavor oils and weighting agents. Flavor oils are usually composed ofessential oils or citrus oils. Weighting agents can also be called density-adjusting agents, because they are added to flavor oils to increase theoil-phase density. They are materials that are oil soluble, which have noflavor of their own and have densities higher than the flavor oils. Forbeverage cloud emulsions, the oil phase contains only flavorless oils andthe weighting agent. The flavorless oil can be orange terpenes or otherflavorless oils, such as vegetable oils or edible waxes. The water phaseusually consists of various types of hydrocolloid, acid, preservative, andcoloring. The most commonly used hydrocolloids are gum arabic. Duringthe shortage of gum arabic in the mid-1980s, blends of different hydro-colloids and modified food starches were developed for uses as gum arabicreplacements.


III. INGREDIENTS

A. Oil Phase

1. Oils

In most of the beverage flavor emulsions, the flavor oils is the major com-ponent of the oil phase. It is responsible for providing the flavor and mayproduce some of the cloud in the beverage. The flavor oil in the emulsions iscomposed of several types of citrus oil in different proportion to producea well-balanced flavor. Citrus oils are characterized by the presence ofmore than 90% of monoterpenes and a smaller amount of sesquiterpenes.They are not miscible with water. To compose a total flavor for a beverage,other flavor chemicals, such as alcohol, aldehyde, ketone, and esters, arecompounded. These minor ingredients are responsible for the characteristicaroma and flavor profiles. These chemicals are important for the flavor ofthe products but have different degrees of solubility in water and do notbehave as the typical oil phase in the emulsion (3).

In the beverage cloud emulsions, because terpenes possess little intrinsicodor or flavor, terpene hydrocarbons are used often as the oil component inthe emulsions. Pure and deodorized vegetable oils are also commonly usedwith terpenes in the cloud emulsions. In addition to vegetable oils, ediblewaxes can also be used in the beverage cloud emulsions. When a strongercloudifying strength is needed in some beverage flavor emulsions, terpene orvegetable oils are added to the flavor oils in order to produce more clouds inthe emulsions (4,5).

Although citrus oils and other oils are important for the beverageemulsions, they also present a major problem for the emulsions because oftheir low density. In general, citrus oils have a specific gravity in the range of0.845 to 0.890 g/cm3. The specific gravity of a 10–12% sugar solution of softdrink is about 1.038–1.046 g/cm3. The low specific gravity and insolubility ofcitrus oils in water indicate that it is difficult to mix the oils into the sugarsolution to make a stable dispersion of the oils. Therefore, weighting agentsare needed to add to the oils to increase the specific gravity.

2. Weighting Agents

Weighting agents are a group ofmaterials added to the essential oils to increasetheir specific gravity and help to maintain a stable dispersion of the oil in sugarsolution. In the beverage industry, weighting agents are also called density-adjusting agents. The criteria for a good weighting agent are as follows:

� Oil soluble, but relatively insoluble in sugar solution� Has specific gravity sufficiently higher than that of essential oils


� Contributes no undesirable flavor, odor, or color to the finishedproduct

� Approved by the food regulatory agency of the country in whichthe product is to be consumed

In the 1940s, brominated vegetable oil (BVO) was first used in softdrinks as the weighting agents for orange oils. In 1970, the United Kingdomand several other European countries withdrew their permission for theuse of BVO. The United States and Canada restricted the permitted levelof use to 15 ppm (6–10). These limitations have given the soft drink industrydifficulty in producing flavor emulsions for soft drinks. Since then, a largevariety of materials have been investigated for use as weighting agents forcitrus oils.

As the result of the limitations placed onBVOuse in 1970, the commonlyused weighting agents are ester gum, sucrose acetate isobutyrate (SAIB), anddamar gum. These materials have been approved individually by variouscountries; however, none has been approved universally by all countries.

Ester gum. Ester gum is a hard, pale-amber-colored resin producedby the esterification of pale wood rosin with food-grade glycerol and purifiedby steam stripping. Wood rosin is a solid resinous material which occursnaturally on the oleoresin of pine trees. There are three major sources forrosin (11):

Gum rosin. Obtained by collection of the exudate of living pine treesfollowed by heating and stripping of the volatile terpenes andturpentine compounds

Wood rosin. Obtained by solvent extraction of aged pine stumpsfollowing by solvent refining of the extracted material

Tall oil. Obtained as a by-product of the Kraft process for woodpulp production followed by depitching and distillation

All three types of rosin are similar in nature and are composedof approximately 90% resin acids and 10% nonacidic neutralcompounds.

The acid fraction is a complex mixture of isomeric diterpenoid mono-carboxylic derivatives of alkylated hydrophenanthrenes having the typicalmolecular formula C20H30O2. These acids are classified into two types: theabietic type and the pimaric type. The distribution of these resin acidscan vary slightly according to botanical origin. Resin acids containingconjugated dienes and dehydro-, dihydro-, and tetrahydroabietic acids areclassified as being of the abietic type. These differ from pimaric type acids inlocation of the double bonds and in type and configuration of the alkylgroups attached to C-13. The two double bonds of the abietic-type acids


are conjugated, whereas those of the pimaric type cannot be because of thequarternary nature of the carbon atom at position 13. The presence of theconjugated double bond of abietic acid makes it readily oxidizable, isomer-ized by heat and acids. The structural formulas of abietic acid and pimaricacid are shown in Fig. 1. A typical analysis of gum and wood resins showsthat the abietic acid content is in the range 55–68%.

In the neutral fractions of gum, wood, and tall oil rosins, one featurein common is that each is composed of approximately 60% of ester of resinand fatty acids. The resin acids are those found in the acid fraction of rosin.The fatty acids are those found in other neutral products. The fatty acids arepredominantly the C-18 acids (i.e., oleic, linoleic, linolenic, and stearic). Thealcohol portion of the esters is unidentified.

Ester gum is made by esterification of wood rosin with glycerol. Thestructurally hindered nature of the resin acid carboxyl group attached to atertiary carbon atom makes it necessary to use higher temperatures or gen-erally more drastic conditions to bring about esterification. This hindranceis, in turn, responsible for the unusual resistance of the ester linkage tocleavage by water, acid, or alkali. The ester gum with optimum physicalproperties is glycerol triabietate, which is produced from 3mol of resin acidcombined with 1mol of glycerol. In commercial practice, it is difficult toobtain such a perfect ester. Commercially, the esterification process pro-duces a mixture of monoglycerides, diglycerides, and triglycerides. Excessglycerol is removed by vacuum distillation and the remaining ester gum issteam-sparged until-odor free. Ester gum produced in this manner has thefollowing typical properties (12):

Softening point (�C) 90Color, USDA rosin scale WGAcid number 6.5Density at 25�C (g/cm3) 1.08

Figure 1 The structural formulas of (a) abietic acid and (b) pimaric acid.


Ester gum is produced in flaked and rock forms in the past. The flakedform is easier to use. Because of the large surface area present on theflaked form, flaked-form resins are also prone to gradual oxidation. Thisusually induces darkening in color and decreasing solubility in terpenes andother organic solvents. In 1997, the manufacturer of ester gum started toproduce the ester gum in pastilles shape and packaged in multilayered bagsto protect the product. It is claimed that the barrier bag protected theproduct properties for at least 8 month (13).

The use of ester gum (glyceryl abietate) in orange oil terpene fractionsto obtain a stabilized orange drink was first proposed in 1967 (14). Estergum is considered the best agent for creating cloudy emulsions when it isused in combination with other emulsifying agents (5). Because ester gum isprone to oxidation, hydrogenated abietic acid and esters were proposed foruse in cloud emulsions but are still pending government approval (15). Gly-cerol ester of wood rosin or ester gum is approved by the United States,Canada, and a number of other countries as a food additive. As specified inthe U.S. Code of Federation Regulations, Title 21, under Section 172.735(Glycerol ester of wood rosin), ester gum is permitted for use in nonalcoholiccarbonated beverages at 100 ppm. Ester gum is known in the EuropeanUnion under food additive number E445. It has been evaluated by theScientific Committee for Food of the EEC, who established an acceptabledaily intake (ADI) of 12.5mg/kg body weight/day. The JECFA (JointExpert Committee on Food Additives) of the FAO/WHO listed ester gumin the codex Alimentary as E445, glycerol ester of wood rosin (16,17).

Sucrose Acetate Isobutyrate. Sucrose acetate isobutyrate (SAIB) is amixture of sucrose esters containing approximately 2mol acetate and 6molisobutyrate per mole of sucrose. The major constituent of the mixture is6,60-diacetyl- 2,3,4,103040-hexaisobutyryl sucrose. The structure formula ofpure sucrose acetate isobutyrate is shown in Fig. 2. SAIB is manufacturedby esterification of sucrose with acetic anhydride and isobutyric anhydridein the presence of a barium hydroxide catalyst. After the removal ofunreacted acids and anhydrides by steam-stripping under vacuum, the crudeproduct is decolorized with activated carbon and purified by moleculardistillation (18).

Sucrose acetate isobutyrate is a tasteless, odorless, and colorlessviscous substance. It has a viscosity of 100,000 cps and a specific gravity of1.146 g/cm3 at 25�C. SAIB is very soluble in various solvents, such as orangeterpenes, ethanol, and median-chain triglycerides. Being fully saturated,it has excellent oxidative stability. The viscosity of SAIB decreases whenit is mixed with a solvent and, similarly, the viscosity will drop from 100,000to 105 cps when the temperature is raised from 30�C to 100�C (19). The typical


properties of SAIB are as follows:

Color, Gardner Scale: 1Refractive index, n 20/D: 1.154Specific gravity, at 25�C (g/cm3) 1.146Flash point, Tag closed cup (�C) 226Solubility in water, at 25�C (wt%) 0.1Shelf life (years) 2

A study conducted by Eastman Chemical Company determined thestability of SAIB in beverages. The study was performed at various valuesfor the pH from 2.46 to 4.01 and for the temperature from 20�C and 40�C.SAIB was very stable under all conditions for 2 months. Minimal degrada-tion of SAIB was observed in the beverages and beverage bases. The studywas concluded. The stability of the emulsion containing acacia gum can beimproved by the addition of six to eight parts per million of dioctyl sodiumsulfosuccinate (DSS), based on the final beverage (20).

Sucrose acetate isobutyrate is metabolized to sucrose, acetic, and iso-butyric acids. Metabolic studies indicate that the primary urinary metabolitesin humans are lower aceylated sucrose and free sucrose. In 1968, a patent wasissued in the United Kingdom for SAIB to be used for mixing with essentialoils to facilitate the emulsification of these oils in water for the manufacture ofsoft drinks (21). Since then, many studies have been made on the biochemicaleffect and metabolism of SAIB in the rat and man. In general, the resultsof the studies show that the use of SAIB as a food additive is unlikely toconstitute a toxicological problem with respect to its metabolitic fate in man(22–24). SAIB is now being used as a weighting agent in beverage emulsions inmany countries. In June 1999, the United States Food Drug Administration

Figure 2 The structural formula of sucrose acetate isobutyrate (SAIB).


(FDA) approved SAIB for use in nonalcoholic beverages at a maximum levelof 300 ppm in the finished beverage (25). SAIB is approved globally in morethan 40 countries. In most of the countries of the European Union, thepermitted level is 300 ppm or 300mg/L. In the United States, SAIB can belisted as SAIB or sucrose acetate isobutyrate. In the European Unioncountries, SAIB is designated as E444 (26).

Damar Gum. Damar is the general name given to a group of naturalexudates from shrubs of the Caesalpinaceae and Dipterocarpaceae familiesand other families belonging to the genera Dammar (27). These plants areindigenous to Malaysia, Indonesia, and the East Indies. The materials of theresins are known as Pale Bold Indonesia Dammar, Dammar Mata Kuching,or Cat Eye Dammar (28).

Damar resin contains an acidic fraction and a neutral fraction. Theneutral fraction of the resin is subdivided into ethanol-soluble and ethanol-insoluble fractions, which are designated a-resene and b-resene, respectively.The b-resene has been shown to be a polymeric material of low molecularweight. The remainder of the resin is principally a complex mixture ofneutral and acidic triterpenes. The main neutral triterpenes are dammadi-ennone, dammadienol, hydroxydammarenones, and hydroxyhopanone.All of these substances are tetracyclic compounds.

In the acidic fraction, the following acids have been identified: dam-marolic acid, ursonic acid, and dammarenolic acid, dammar-enonic acid.Methyl esters mixture of the above acids are found with the acids (28,29).

Damar gum is completely soluble in essential oils, benzene, toluene,petroleum ether, carbon tetrachloride, and chloroform. It is partially solublein ethanol and acetone, and it is insoluble in water. It has been used incoating varnishes for many years. Because of its high solubility in essentialoils, it is used as a weighting agent for edible oil to produce beverage cloudsemulsions (30). Damar gum can be used in the crude form if it is of topquality; small amounts of impurities that consist of insoluble materials, suchas fine bark chips, can be removed by simple filtration after the solution ismade. However, in most cases, the crude form will impart a terpene note tothe product. A deodorized material can be made through solvent extraction(28,29,31) or by fractional distillation (32).

The typical characteristics of damar gum for use as weighting agentare as follows (27):

Appearance: Off-white to pale yellow to brownWeight loss: 5% at 105�C for 18 hMelting point range: 90–110�CInsoluble matters: 0.5%Specific gravity: 1.05–1.08 g/cm3at 20�C


Damar gum is permitted for use in beverages as an all-natural non-chemically modified vegetable gum resin; it is classified under the nameAGATIS D’AMMARA (Classification N3), number 16 of the Single Listof Natural Flavoring Substances, countersigned by 21 countries belongingto the Council of Europe.

Brominated Vegetable Oils. Brominated vegetable oil (BVO) was firstused as the weighting agent for essential oils in beverages in the 1940s. It ismade by the addition of bromine molecules to the olefinic bonds of theunsaturated fatty acid moieties of vegetable oils.

Commercial BVOs are made from vegetable oils with an iodinenumber of 80–90, such as olive oil, which produces a BVO with a specificgravity about 1.24 g/cm3 at 20�C. Vegetable oils with an iodine number of105–125, such as sesame oil, corn oil, soybean oil, or cottonseed oil, give aBVO with a specific gravity of about 1.33 g/cm3 at 20�C. The BVO forbeverage use is a dark brown color, viscous liquid with bland odor andbland taste.

It is the high specific gravity of BVO that makes it unique as animportant weighting agent for citrus oils in soft drinks. It balances thespecific gravity of citrus oils and, at the same time, provides cloudiness oropacity to the beverages. In 1966, the Joint FAO/WHO Expert Committeeon Food Additives expressed particular concern about the possibility ofbromine storage in body tissue when BVOs were used in food. In 1970,permission to use BVOs in soft drinks was withdrawn in the UnitedKingdom, after being permitted for use for some 30 years. In the sameyear, the U.S. government limited the use of BVO to 15 ppm in the finishedbeverage (6).

Other Weighting Agents. Since the use of brominated vegetable oilwas banned or limited in different countries, the soft drinks industry has beenfaced with the problem of finding alternatives to replace BVO. In additionto ester gum, SAIB, and damar gum mentioned earlier, the followingpotential weighting agents have been proposed and are still being evaluatedby the soft drink industry and metabolism and toxicology studies (4):

Sucrose octa-isobutyrateSucrose octa-acetateSucrose hepta-isobutyrateSucrose octa-propionatePropylene glycol dibenzoateGlycerol tribenzoateGlycerol ester of hydrogenated rosinMethyl ester of hydrogenated rosin


B. Water Phase

1. Water

In beverage emulsions, water is the major component. In most beverageemulsions, the water content is about 60–70%, and in certain formulations,it can be as high as 80%. The importance of good quality water in anemulsion is no less than that in soft drinks. Standard water treatment forsoft drink water should be applied to the water intended for beverageemulsion. The treatment should remove colloidal and suspended matter,undesirable taste, odor, and micro-organisms. The carbonate hardness oralkalinity of water should be reduced. Highly alkaline water neutralizes theacid in the soft drink and causes the beverage to ‘‘go flat’’ and taste insipid.It will also affect the stability of the beverage emulsion due to the neutral-ization of the electrostatic charge carried by the emulsion particles. Theeffect of high water hardness or alkalinity on emulsion stability is muchmore obvious in the finished soft drink than in the emulsion concentratebecause of the very high volume ratio of water to emulsion in the finisheddrink. Although there is no industrial standard set for hardness or alkalinityfor water used in soft drinks, the major soft drink companies have beenusing the maximum hardness level at 50mg of CaCO3 per liter for coladrinks and 100mg of CaCO3 per liter for other products (33,34). Forbeverage emulsions, the water hardness level is recommended to have thesame quality—not to exceed 50mg of CaCO3 per liter. Public municipalwater supplies are not necessarily acceptable for soft drinks or emulsionswithout treatment. The water treatment at the public water plant is toproduce safe potable water and it does not necessarily have the qualityrequired for the soft drink industry. Public water companies depend onreservoirs for water supplies. Water quality varies, because differentwater sources feed into the reservoirs. Seasonal variation is anotherinfluence on the quality of the water in the reservoirs. The additionaltreatment applied to municipal water at the soft drink plant is to ensurethat the quality of water used throughout the year is the same. In thepreparation of beverage emulsion, the following steps further treat theraw municipal water: cartridge filtration (5 mm), cation exchange, anionexchange, and ultrafiltration (0.22 mm). This water must be of such qualitydescribed as ‘‘ultrapure’’ (35).

2. Hydrocolloids

Hydrocolloids are water-soluble biopolymers consisting of high-molecular-weight polysaccharides with rigid backbone and totally hydrophilic polyol


(sugar) moieties. Some hydrocolloids serve as the stabilizer in theoil-in-water emulsions (36). The basic mechanisms for emulsion stabilizationby hydrocolloids are viscosity effect, film formation, steric hindrance, andelectrostatic interaction. These stabilization mechanisms will be discussedin detail later. To perform as an effective stabilizer for beverage emulsion,the hydrocolloid must have the following properties:

� Readily soluble with high solubility in cold water� Low viscosity in water� High emulsifying property� Will not thicken or gel on aging

Gum Arabic. Gum arabic or gum acacia is the most well-knownhydrocolloid for use in beverage emulsions. It is a dried exudate from thestems and branches of trees of the genus Acacia, which belong to the bota-nical family Leguminosae. About 80% of the acacia of commerce are fromAcacia senegal, with Acacia seyal providing about 10%, with other minorspecies making up the difference. The best commercial grades designated forgood uses are those giving clear, viscous, colorless, or pale yellow solutionsthat are as tasteless and odorless as possible (37). Gum arabic is a verycomplex substance with many unique features and properties. Supposedly,its chemical structure is composed of a main skeleton of 18–20 galactoseunits, with arabinose and rhamanose units limited to the side chains anduronic residues existing as sodium, potassium, and magnesium salts, whichtend to give the molecule a buffering capability. The geometry is such thataround 50% of these units are believed to be internal in the molecule (38).Composition ranges for the more commonly occurring Acacia species are asfollows (37):

Ash 3–4 %Nitrogen 0.14–1.11%Glucuronic acid 9–16%pH, solution at 25% 4.4Intrinsic viscosity 12.1–20.7ml/gMolecular weight (312–950)� 103

Gum arabic is almost completely soluble in twice its weight of waterand has very low viscosity in water with regard to concentration. A 30%solution has a viscosity of about 100 cps (39,40). Solutions of gum arabicshow essentially Newtonian flow properties below about 40% concentration(41). The pH of a 10% gum solution is about 4.6–5.5. The addition of saltsor electrolytes affects the consistency of the gum solution as does the pH.Gum arabic solutions become most viscous near pH 6–7. The high solubility


and low viscosity allow for the preparation of solutions containing a highconcentration of gum solids.

Gum arabic is well known as an emulsion stabilizer and emulsifier.It is most effective in stabilizing oil-in-water emulsions. It has beenreported that gum arabic has a hydrophile–lipophile balance (HLB)value of 8.0 (42); another researcher gave it another value of 11.9 (43).These values indicate that gum arabic is an efficient emulsifier for oil-in-water emulsions. Gum arabic reduces the interfacial tension between oiland water and facilitates the formation of fine oil droplets in the emulsion(94). It has the ability to form an adsorbed film at the oil–water interfacewhose surface viscoelasticity is rather insensitive to dilution of the aqueousphase. The formation of a thick layer of film around emulsion dropletsenable the flavor oil emulsion to be sterically stabilized both in a concen-trate and in a diluted beverage (44). It has been reported that it is theprotein-containing high-molecular weight-fraction which adsorbs moststrongly at the oil–water interface and is probably mainly responsiblefor the emulsifying and stabilizing properties of the gum acacia (45). Ina functionality study of fractionated gum arabic, it was reported that thehigher-protein, high-molecular-weight fraction makes the best emulsions.All fractions except the lowest-molecular-weight fraction (about 5% of thewhole gum) make reasonably good emulsions and perform well in modelbeverage (46).

Spray-dried gum arabic is a cleaned gum in a spray-dried powderform. It is used commonly for the preparation of beverage emulsions. Inproducing spray-dried gum, the raw gum is crushed, dissolved in water,foreign materials of the gum in solution removed by filtration or centri-fugation, and then pasteurized and spray-dried into powder. A standardspray-dried gum arabic contains approximately 50% of particles less than75 mm. A recent development in the gum industry is to make the powderinto a fine granulated form for dust-free flowing and faster dissolution inwater.

Modified Food Starches. The most widely accepted alternative togum arabic for use as a beverage emulsion stabilizer is modified starches.They are a group of specially designed starch derivatives with balancedlipophilic and hydrophilic groups on the starch molecules. The starch deri-vatives are also converted to low-viscosity starches by acid degradation orby enzyme digestion. Caldwell and Wurzburg (47) and Richard and Bauer(48) developed the modification processes. The starch derivative is preparedby a standard esterification reaction in which the reagent and thestarch suspended in water are mixed under alkaline conditions. The reagentis a substituted cyclic dicarboxylic acid anhydride having the following


structural formula:

where R is a dimethylene or trimethylene radical and R0 is the substituentgroup, usually a long hydrocarbon chain. Examples are the substitutedsuccinic acid anhydrides in which the substituent lipophilic chain is analkyl or alkenyl group containing 5–18 carbon atoms. The acid ester ofstarch may be represented by the following structural formula:

where R is a dimethylene or trimethylene radical and R0 is the substituentlipophilic group.

The octenylsuccinic acid anhydride-treated starches, generally referredto as sodium starch octenyl succinates, have been approved for food use bythe U.S. FDA and by several other countries. The maximum level of theanhydride treatment allowed is 3%. The approximate degree of substitu-tion is about 0.02 (49). These starches are widely used in the food andpharmaceutical products.

This type of modified starch is a white or yellowish-white bland-tasting, odorless powder. It is cold water soluble and the solution is very lowin viscosity. At a concentration less than 20%, its solution has a viscosityabout equal to that of gum arabic at the same concentration. The lowviscosity of this starch in solution is an important attribute for it to beused as a beverage emulsion stabilizer. Another important property is theemulsifying power derived from the lipophilic modification of the starchmolecules. In comparing the use of modified starch to gum arabic, modifiedstarch is a cleaner material, as it contains fewer foreign substances. It is aseasy to handle as gum arabic. The emulsion made with some of the earlierdeveloped modified starch had the tendency to develop higher viscosity orform gel in storage at refrigerated temperatures due to retrogradation of thestarch. This shortcoming has been overcome by recent developments(50,51).


There are several advantages to using modified starch as the stabilizerin the emulsions. The most important one is that the raw material formanufacturing this starch is corn, which is grown in several areas ofthe world and the overall supply is less likely to be affected by climaticconditions than that of gum arabic. The second is that in formulation, asmaller amount of modified starch than gum arabic can be used to stabilizethe emulsions. A drawback of using modified starch is that it cannot beconsidered as natural an ingredient as gum arabic.

Gum Tragacanth. Gum tragacanth is known to have the excellentstabilization mechanisms of colloidal suspensions even at relatively lowconcentrations (of the order of 10 ppm). Gum tragacanth is commonlyused with gum arabic as the stabilizer for emulsions because of its highviscosity. When it is used in conjunction with gum arabic, the mixture is avery effective stabilizer for emulsions. The viscosity of a solution of gumtragacanth and gum arabic tends to be lower than that of either constituentsolution. A minimum viscosity is attained in a mixture consisting of 80%tragacanth and 20% arabic (52). The viscosity of gum tragacanth is moststable at pH 4 to 8 (53). In solution, it takes 48 h to develop to its maximumviscosity. The viscosity of the tragacanth solution is reduced by the additionof acids, alkalis, or electrolytes. It is also affected by the manner in which thesolution is prepared.

Gum tragacanth is not a completely soluble gum. It is compatible withother plant hydrocolloids as well as carbohydrates and proteins. Because ofthe high price of this gum and its slow hydration rate with slow viscositydevelopment, it is not commonly used commercially in beverage emulsions.

Other Hydrocolloids. Other hydrocolloids which have been reportedto be useful in beverage emulsions are some ‘‘underutilized’’ species ofAcacia, such as Acacia verek and gum arabic from the Acacia senegalspecies, and locust bean gum (54). The successful performance of thesegums depends very much on their formulations. There are reports onusing hydrocolloid materials such as propylene glycol alginate, xanthangum, pectin, gellan gum, ghatty gum, and carboxymethylcellulose toenhance the stability of flavor and cloud emulsions. These materials areused as a thickener or blend of thickeners in diluted beverages. Theycannot be used to replace gum arabic or modified starch in the emulsionconcentrate. The beverage emulsions prepared using these hydrocolloidmaterials do not have the required stability for the emulsion concentrateand in the finished soft drinks. In general, polysaccharides are usually addedto oil-in-water emulsions to enhance the viscosity of the aqueous phase,which produces desirable textural characteristics and retards the creamingof oil droplets. Nevertheless, at certain concentrations, polysaccharides


have been shown to accelerate creaming instability by promoting dropletflocculation through the depletion mechanism (55,56).

3. Acids

Acid is an important ingredient in a soft drink because it provides the tasteand controls the pH. In a beverage emulsion, acid plays an important role incontrolling the pH to prevent the growth of micro-organisms. It is used tobring the emulsion to pH below 4.5. The reason is that most bacteria growbest at a pH range of 6.0–8.5, and pathogens do not grow well below pH 4.5.

Citric acid is used most commonly in beverage emulsions because it isclosely related to citrus products and citrus flavor is the most popular flavorof beverage emulsions. There is an additional benefit of using citric acidbecause its sequestering power for chelating metallic ions may be presentin water (57). Citric acid is usually produced in crystal or powder form andis readily soluble in water. Other acids, such as phosphoric acid, is used incola emulsion but is not permitted in beverages which claim a fruit juicecontent. Malic, tartaric, acetic, and lactic acids have been used somewhatrarely as a replacement for citric acid for producing acidity.

4. Preservatives

Benzoic acid or sodium benzoate is added to beverage emulsion as a pre-servative. Other approved preservatives can also be used. Because of the lowsolubility of benzoic acid, its sodium or potassium salt is most commonlyused. The preservative effect of benzoic acid is largely influenced by the pH.It is greatly increased by a corresponding decrease in pH because it is theundissociated form of benzoic acid which exhibits the preservative action.The preservation of beverages by benzoic is most effective when acidity isless than pH 4.5.

5. Colorings

FD&C colors are commonly used for citrus flavor emulsions. FD&C Yellow6 and FD&C Red 40 are used for an orange shade. Natural colors, such as�- and �-carotenes and marigold extracts, are used for their yellow ororange shades. However, the stability of natural colors is never as good asthe FD&C colors. In cola-flavored soft drinks, caramel color is used.Because of the high acidity in the beverage, the acid-stable double-strengthcaramel color is used. When the emulsion is a cloudifier, usually no colorsare added. The natural color of cloudifiers is just milky white. Becausetitanium dioxide is an acceptable food additive, the water-dispersible formof it has been used to improve the opaque appearance of oil-in-water


emulsion for short-term stability (58,59). Because of the high specific gravityof titanium dioxide it has the tendency to settle out in the beverage.

III. PREPARATION OF BEVERAGE EMULSIONS

In general, the preparation of a beverage emulsion can be divided into thefollowing steps:

A. Step 1. Preparation of the Water Phase and Oil Phase

For the water phase, it is simple to dissolve the proper amounts of preser-vative, citric acid, coloring, and gum in water and make a complete solution.It is important to follow the proper order of the addition of these ingredientsto assure their full dissolution in water. For the oil phase, the weightingagent is dissolved completely in the oil. The ratio of the weighting agent tothe oil is governed by the legitimate permissible amount in the finishedbeverages. Usually, the maximum quantity of the permissible amount ofweighting agent is used to fully utilize the benefit of the weighting agent.

B. Step 2. Prehomogenization

In this step, the oil phase is mixed with the water phase to make a crudeemulsion or ‘‘premix.’’ It breaks the oil phase into small oil droplets in thewater phase. The quality of the premix of prehomogenization being suppliedto the homogenizer can greatly influence the quality of the finishedemulsion. A good premix fed to the homogenizer produces a better emulsionthan a poor premix (60). The reason for this is quite easy to understand.At any given homogenizing pressure, a fixed amount of energy is availablefor transmission to the product; this energy is the means by which particlesize reduction is achieved. If a significant portion of this energy is needed toreduce very large particles, then there will not be enough energy left over towork on the smaller particles and reduce them even further.

As a general rule, the premix should have a droplet size less than 20 mmto avoid a polydispersed final emulsion (61). In practice, it is preferable tohave the premix droplet size smaller than 10 mm. Usually, this step can beachieved by the use of a high-speed mixer, colloid mill, homomixer, ormaking one pass through the homogenizer.


C. Step 3. Homogenization

This is the most important step of the process. In this step, the crudeemulsion or premix is pumped through the homogenization valves of thehomogenizer at high pressure. The high pressure forces the liquid topass through the valves at a high velocity, which creates turbulence andcavitation forces that shatter the oil droplets into fine particles (62,63).Single-stage or two-stage homogenizers are usually used for this procedure.For beverage emulsions, a two-stage homogenizer is preferred. In a two-stage homogenizer, the second stage provides controlled back-pressure toensure the optimum efficiency of homogenization. An additional benefit ofthe second-stage valve is its ability to delay or prevent reagglomeration of theparticles after leaving the first stage. The second-stage valve also provides ameans of controlling the viscosity of the product (64). The pressure settingfor the first-stage homogenization valve usually varies from one emulsion tothe other and is dependent on the composition of the emulsion. This first-stage pressure can vary from 2000 psi to 5000 psi, or from 140 to 350 kg/cm2,and in certain cases, 7000 psig or 500 kg/cm2. In general, the second-stagepressure is set between 300 and 500 psi (20 and 35 kg/cm2) or set at 10% ofthe first-stage pressure.

Because beverage emulsions usually require either a very small averageparticle size and a very uniform particle size distribution, it is simply notpossible to reach these goals in a single pass through a homogenizer.Generally, two passes of the emulsion through the homogenizer areperformed in order to obtain a more uniform particle size distribution inthe emulsion. Because beverage emulsions vary in their formulations, thereis no set operational procedure for all emulsion preparations. However, theideal operating procedures for each emulsion can be designed accordingto the basic principles of a stable emulsion.

IV. STABILITY PROBLEMS

In 1970, the use of BVO in soft drinks has been banned or regulated. The useof BVO in the soft drinks is limited to 15 ppm in the United States and a fewother countries. In many countries, its use is completely prohibited. For thisreason, the soft drink industry has been trying to find other appropriatematerials to take the place of BVO for use as the weighting agent for citrusoils in beverage emulsions. Although ester gum, SAIB and damar gumcan be used as the weighting agents in place of BVO, they have their short-comings and limitations. All of these weighting agents do not have aspecific gravity as high as that of BVO. The specific gravity of ester gum


is 1.08 g/cm3, that for SAIB is 1.15 g/cm3, and that for damar gum is1.05 g/cm3. They all have government regulations on the amount can beused in beverages. Under these regulations and the low specific gravity ofweighting agents, a stable beverage emulsion has become difficult to makethan when BVO could be used without limitation.

The instability of beverage emulsion observed in both the concentratesand the finished soft drinks may lead to the following occurrences (65):creaming (ringing), flocculation, and coalescence.

A. Creaming

Creaming is a term adopted from the separation of cream from unhomo-genized milk. In the soft drink industry, the common term for creaming ofthe soft drink in bottles is ‘‘ringing.’’ This is because in the beverages, theflavor oil particles separate from the beverage and float to the top and it isseen as a white creamy ring at the neck of the bottle. Creaming or ringing isrelated to flocculation. It can be considered as a separation of one emulsioninto two emulsion sections. One section is richer in the oil phase than theoriginal emulsion, and the other section is richer in the water phase thanthe original emulsion. When creaming occurs in a bottle of beverage, theemulsion richer in the oil phase rises to the top and forms a creamy layer orjust a ring at the neck of the bottle. It is not only unsightly to have the ringat the neck of the bottle but it indicates the breakdown of the distribution ofthe flavor oil in the bottle of beverage.

In relation to creaming, two other phenomena not commonlyobserved in bottles of soft drink are ‘‘lifting’’ and ‘‘striation.’’ Liftingoccurs when the emulsion in the bottle of beverage lifts up from thebottom and shows a clear layer of liquid at the bottom. Striation occurswhen the emulsion in the bottle shows two or more distinctive layers ofdifferent degrees of cloudiness. It looks as if the emulsion in the bottlehas separated itself into two or more different particle size or densityfractions. Lifting and striation only show in bottles that have been held inan upright position undisturbed for a long period of time. They are moreoften seen in bottles that are kept in the refrigerator where the temperatureis low and the Brownian movement and thermal convection of the emulsionparticles are less active.

A rather special and rare situation found in soft drinks is ‘‘sedimentation.’’Sedimentation is also called ‘‘downward creaming’’ by the soft drinkindustry. It happens only when the weighting agent is overused or super-saturated in the oil. In such case, the excess weighting agent will separategradually from the flavor oil and precipitate in the beverages (66). When the


gum arabic or starch used in preparing the emulsion has not been purifiedproperly, the impurities or foreign materials will precipitate as sediments.

B. Flocculation

Flocculation occurs when oil droplets of the dispersed phase formaggregates or clusters without coalescence. At this stage, the droplets stillretain their original identities. The forces which draw these droplets togetherto form aggregates are primarily the long-range London–van der Waalsforces and electrostatic forces around the droplets (67). From the creamingpoint of view, these aggregates behave as simple large droplets. The rate ofcreaming is accelerated in systems in which the density difference of theaggregates from the continuous phase is sufficiently large. In the emulsionconcentrate, a perceptible increase in emulsion viscosity can be observedwhen flocculation occurs. Although flocculation generally changes thephysical properties of the emulsion, the particle size distribution remainsunchanged. In the finished beverage system, the droplet concentration is solow that the flocculation is often reversible. The aggregates can be readilyredispersed because the interaction forces between the droplets are weak.This phenomenon can be observed by lightly shaking a bottle of beveragethat has a ring in the neck. The ring quickly disappears after the shaking.

C. Coalescence

In this stage, there is localized disruption of the sheaths around neighboringdroplets of the aggregates, and the oil droplets merge together to form alarge droplet. This leads to a decrease of the number of oil droplets andeventually causes the breakdown of the emulsion. When a proper hydro-colloid is used in the water phase, the breakdown of the emulsion willseldom reach this stage. The reason is that the hydrocolloid, such as gumarabic, has a good film-forming ability and will form film around the oildroplets in addition to providing viscosity in the water phase (68). Moreabout the film-forming property of gum arable will be discussed later in thischapter.

V. STABILIZATION OF BEVERAGE EMULSIONS

For a beverage emulsion, the most critical criterion of stability is its stabilityin the finished beverage, where the emulsion concentrate is further dispersedin sugar solution. The stability of the emulsion in the concentrate is mucheasier to achieve than in the finished beverage. This is because, in the


concentrate, the viscosity is high due to the high concentration of a hydro-colloid, which acts as a stabilizer. In the beverage, the emulsion concentrateis redispersed in sugar solution at a very high dispersion ratio. Thedispersion ratio could vary from 1/300 to 1/1000 depending on the flavorstrength or the cloud strength in the emulsion concentrate. It can almost bedescribed as the emulsion concentrate being dispersed in a second waterphase i.e., from a gum-solution water phase to a sugar-solution waterphase. The following discussion will emphasize the principles involved instabilizing beverage emulsions in sugar solutions such as ready-to-drinkbeverages.

A. Stokes’ Law

The ‘‘Ringing Test’’ is the most popular method used to evaluate thestability of beverage emulsions in soft drinks. It is a simple test in whichbottles of soft drinks containing the beverage emulsion are held in anupright or horizontal position for observation of ringing. If the emulsionin the soft drink bottle is not stable and will ring, it will form ring faster inthe horizontal position than at the upright position. The reason is simplythat the oil droplets in the bottle have a much longer distance to travel to thetop surface than in the bottle at horizontal position. The rate of ringing orcreaming of an oil droplet in sugar solution may be determined by equatingthe force of gravitation with the opposing hydrodynamic force as given byStokes’ law:

� ¼2gr2ð�2 � �1Þ

9�1ð1Þ

In Eq. (1), � is the rate of creaming or sedimentation, g is the accelerationof gravity, r is the droplet radius, �2 is the density of the oil phase, �1 isthe density of the water phase, and �1 is the viscosity of the water phase.In an o/w emulsion or a soft drink, the oil density, �2, is lower than that ofthe water phase, �1. The resulting sign, �, is negative; hence, creaming orringing will occur.

As an example of the use of Stokes’ law, consider the case of theorange flavor emulsion used to make a beverage. Orange oils are themajor components of orange flavor. Typically, orange oils have a densityof 0.846 g/cm3 and the sugar solution in the soft drink has a density of1.040 g/cm3 for a 10% sugar solution or 1.048 g/cm3 for a 12% sugarsolution. Applying these density data to Stokes’ law, the resulting � carriesa negative sign, which indicates that the emulsion will ring in the bottle.Stoke’s law shows that the velocity of a droplet, �, is directly proportional to


the density difference between the oil phase and the water phase and to thesquare of the radius of the droplet. It is also inversely proportional to theviscosity of the water phase. The equation clearly shows that if one canmake orange oils with a density equal to that of the sugar solution, therewill be no creaming or ringing in the finished beverage because when �2¼ �1,�2� �1¼ 0 and, therefore, �¼ 0. Because orange oil is lighter in density thanthe sugar solution in the soft drink, weighting agents must be added toorange oil to increase the density. However, the government regulates theuse of weighting agents. These regulations make it impossible to adjust thedensity of the orange oil to equal to that of the sugar solution. For example,an equal weight of ester gum added to orange oil can only bring the densityof orange oil from 0.85 to 0.95 g/cm3. A density of 0.95 g/cm3 is still a largedifference from 1.05 g/cm, the density of a 12% sugar solution. Accordingto Stokes’ law, an emulsion with these components will cause a ringingproblem in the soft drink. Because of the consumer’s preference for thestrength of the orange flavor and the 10–12% sugar sweetness level in thebeverages, there is little a beverage processor can do to narrow the densitydifference between the oil phase and the water phase, because the amountof weighting agent added to the orange oil is reaching the regulated limit inthe finished beverage. The exception is in making diet drinks, whereartificial sweeteners are used and the density of the water solution is almostequal to 1.

As shown in Stokes’ law, if �2� �1 and �1 equal to constants, wheregravity, g, is also a constant, the velocity of an oil droplet moving upwardwill be in direct proportion to r2. It demonstrates that reducing the dropletsize is one effective way to control the creaming velocity when the adjust-ment of oil density has reached the maximum limit permitted. For instance,in a bottle of beverage, a particle 0.1mm in diameter will travel upward at avelocity 100 times slower than a particle 1.0mm in diameter. This exampleshows the importance of controlling the particle size, keeping it small.

B. Adsorption at Interfaces

Most beverage emulsions are made with a water phase composed of gumarabic or other hydrocolloids and water. It has been known for many yearsthat gum arabic solution produces a film at the oil–water interface. Shottonand White reported that gum arabic solution formed a film on paraffin oiland showed this interfacial film to be viscoelastic (68). Dickinson et al.reported that gum arabic forms film and stabilizes an oil-in-water emulsioncontaining gum arabic and n-tetradecane. They also reported that thesurface rheology of gum arabic films is relatively insensitive to dilution ofthe aqueous phase (69). In the case of modified starch, such as sodium


starch octenyl succinates, because it contains both lipophilic and hydrophilicgroups, the starch molecules are attracted to the oil–water interface andform a film about the oil droplets (49).

Because citrus oils are most commonly used in beverage emulsions, weconducted a study of the characteristics of films formed at the interface oforange oil droplets dispersed in a gum arabic solution. In order to simulateclosely to the conditions of commercial beverages, the orange oil wasweighted with ester gum to a density of 0.95 g/cm. The study was repeatedusing modified starch, sodium starch octenyl succinates, in the water phase.The films formed by both gum arabic and modified starch are viscoelastic.Under the microscope, they appeared as elastic interfacial film or shield onthe oil droplets. The elastic property of the films formed by gum arabic andmodified starch on orange oil droplets are seen on the shrunk bubble of oildroplet shown in Figs. 3 and 4 (70). The film or shield protects dropletsfrom coalescence when they collide with each other. Aging of the interfacialfilm in the gum or starch solution seemed to strengthen the film on thedroplets (68,70).

Finished beverages were prepared using emulsions containing orangeoil with ester gum as the weighting agent with gum arabic or modified starchaccording to the standard formula and procedure. The aging study of theseemulsions in the finished beverage was conducted by storing them on a shelfat ambient temperature for 6 months with weekly checking of their particlesize changes. The results show that there was no ringing in the bottle with nosignificant change in particle size as analyzed by Coulter Model LS-130particle analyzer (70). In this study, the orange oil/gum arabic emulsion inthe beverage had a mean particle size of 0.364 mm when freshly prepared and0.410 mm after being aged for 6 months. The orange oil/modified starch

Figure 3 Orange oil droplet in gum arabic solution: (a) full oil droplet; (b) oil

droplet shown with wrinkled membrane after oil had been partially withdrawn.

(From Ref. 70 with permission.)


emulsion in the beverage had a mean particle size of 0.264 mm when freshlyprepared and 0.320 mm after 6 months of aging. The changes are so smalland the effect to the emulsion stability in beverage is negligible. The histo-grams of these two emulsions in beverages when freshly prepared and after6 months in storage are shown in Figs. 5 and 6. These histograms show thatthere is no significant change in the particle size distribution patterns.

Because gum arabic or modified starch is heavier than water, a layer ofhydrated gum arabic or starch will provide an added weight to the oildroplet and actually change the density of the total droplet—the oil dropletplus the hydrocolloid film—to a higher value. When an oil droplet issmaller, the percentage of weight contributed by the hydrocolloid layer tothe total droplet weight will be larger than that for a larger droplet. Thisfurther indicates that to achieve emulsion stability in the finished beverage,the oil droplets of the emulsion should be made small. When oil droplets aresmall, they will gain more of the additional weight benefit contributed bythe gum or starch layer on the particles. This theory answers partially thequestion commonly asked in the industry: ‘‘Why in orange oil weighted witha weighting agent to a density of only about 0.95 g/cm3 could the emulsion

Figure 4 Orange oil droplet in sodium octenyl succinated starch solution: (a) full

oil droplet; (b) oil droplet shown with wrinkled membrane after oil had been

partially withdrawn; (c) orange oil droplets with membrane aged in sodium

succinated starch solution and showed no coalescence. (From Ref. 70.)


Figure 5 Particle size distribution histogram of an orange oil/gum arabic

emulsion in beverage stored at room temperature: (a) 2 days old and (b) 6 months

old. (From Ref. 70.)


Figure 6 Particle size distribution histogram of an orange oil/modified starch

emulsion in beverage stored at room temperature: (a) 2 days old and (b) 6 months

old. (From Ref. 70.)


still be made stable in a solution of density of 1.05 g/cm3?’’ This is becausethe additional weight of the film brings up the density of the particle close tothat of the sugar solution when the particles are small enough to take theadvantage of the weight of film.

The formation of the interfacial film by gum arabic or other hydro-colloid polymers on the oil droplets also helps to stabilize the emulsion inanother way. It is the hydrocolloid material adsorbed on the surface of theoil droplets which prevents oil droplets from coalescence and the formationof larger droplets. Coalescence may eventually lead to emulsion breakdown.It may be described as the adsorbed hydrocolloid layer on droplets keepingthe droplets far enough apart such that the van der Waals attraction force isminimized (67,71). In this way, the droplets will remain dispersed. When anemulsion is stabilized sterically by adsorbed polymers, the mechanism isconsidered ‘‘polymeric steric stabilization’’ (72,73).

In many classic emulsions in which emulsifiers or surfactants are used,the emulsifiers are adsorbed on the interface of oil droplets as a closelypacked monomolecular film and reduce the surface tension. In citrus-flavorbeverage emulsions, Walford proposed using a blend of Atmos 300/Tween-80 (HLB 11.5) to prepare a stable clouding agent (74). Kaufmanand Garti studied the use of a different combination of Span/Tweenemulsifiers to stabilize oil-in-water emulsions (75). They reported thatthe type of emulsifier, the required HLB, oil concentration, amountof emulsifier, and the method of preparation affect emulsion stability.However, because of flavor and their non-natural status, little or noemulsifier is being used in the present commercial beverage products. Incitrus-flavor soft drinks, which have a delicate flavor, the off-flavorimparted from the emulsifier can be detrimental to taste.

In the beverage emulsion, when gum arabic or sodium starch octenylsuccinate is used, they perform as an emulsifier but in a weaker manneras compared to true emulsifiers. It is because these hydrocolloids do nothave the distinctive polar and nonpolar groups that, the real emulsifiershave (42,43,49).

C. Electrostatic Interaction

In an oil-in-water emulsion, the oil particle may acquire an electric chargethrough the ionization of an adsorbed surface charged group. In beverageemulsions, gum arabic is used in the water phase. Gum arabic is an acidicpolysaccharide. The carboxyl (COO�) ions are at the periphery of themolecule and are very active in creating an anionic environment (76).Surface charge may also be acquired through the adsorption of the dissolu-tion of small ions in the water phase. Anions have a greater tendency to be


adsorbed than cations, because cations are normally more hydrated andprefer to stay in the aqueous bulk solution. The electrical charge may alsobe acquired through a possible friction mechanism. An empirical physicalrule may be applied to this theory. It states that a substance having ahigh dielectric constant is positively charged when in contact with anothersubstance that has a lower dielectric constant (77). Because water has adielectric constant higher than oil, oil droplets will have a negative charge.These charges cause a repulsive force among the oil droplets which preventsflocculation, therefore contributing to the stability of emulsion.

As mentioned earlier, dispersed oil droplets may acquire an electriccharge through the ionization of surface groups. Oppositely charged ions(counterions) are preferentially attracted toward the surface, and ions of thesame charge (co-ions) are repelled from the surface. The region of unequalcounterion and co-ion concentrations near the charged surface is called theelectrical double-layer. The double layer may be regarded as consisting oftwo regions: an inner region of strongly adsorbed ions and an outer regionwhere ions are diffusely distributed according to a balance between electricalforces and random thermal motion (78). At some point in the double-layerregion, corresponding more or less to the potential at the zone of shear,the electrical potential is called the zeta potential (83). The determination ofthe zeta potential is important in the study of emulsion stability. It is animportant parameter for both achieving emulsion stability and destroyingemulsion stability. Gum arabic solutions generally have a zeta potentialabout �23mV (71,79,80). However, one cannot categorically state that anemulsion will or will not be stable at a given zeta potential, as some otherfactors should be taken into consideration. It is especially true in the case ofbeverage emulsion; two other important factors are the density differencebetween the two phases and the droplet sizes. It should be stressed that thezeta potential reflects both the electrolyte’s presence in the system and thedissociated ions accompanying the original colloid particles. When a cationelectrolyte is added to an emulsion containing a dispersed phase carryingnegative charges, the electrolyte will be adsorbed and neutralize the zetapotential. This, in turn, will cause aggregation to occur due to theLondon–van der Waal’s forces. This electrolyte effect on emulsion stabilityis much more evident in soft drinks than in the emulsion concentrates,because the emulsion concentrate contains a high concentration of gumarabic and gum arabic is known to produce stable emulsions over a widerange of pH and in the presence of electrolytes (68,70). In soft drinks, theemulsion is in a very diluted form and the gum arabic concentration is verylow; therefore, the electrolyte effect is more obvious.

In the adsorption and desorption of electrolyte ions causing thechange of zeta potential, the higher the valency of the ions, the greater


the effect of compressing the electrical double layer around the droplets andchanging the zeta potential (81); that is, the higher the cation valence, thelower the zeta potential, and the higher the ion concentration, the greaterthe decrease in the negative zeta potential (82). Trivalent ions, such asaluminum ions, have 10–100 times the effect of equivalent concentrationsof divalent ions, like calcium, which have 10–100 times the effect of equiva-lent concentrations of univalent ions. The effect of different electrolytes onthe zeta potential of a colloidal dispersion is shown in Figure 7 (83). For theabove reason, the soft drink industry usually uses only treated water tomake emulsions and the finished beverages.

VI. PARTICLE SIZE CHARACTERIZATION

To achieve good emulsion stability, it is important to control the oil dropletsize. This is especially true for beverage emulsions because their particleshave to be stable in both the concentrate and the diluted finished soft drink.This means that the same oil phase must be stable in two water phases ofdifferent composition. In controlling the particle size, it is essential to knowthe average size of the oil phase particles and the size distribution of these

Figure 7 The effect of electrolytes on the zeta potential of a colloidal dispersion.

(Courtesy of Zeta-Meter, Inc., New York.)


particles in the emulsions. Two emulsions may have the same averageparticle size and yet have dissimilar stability because of the difference intheir distributions of particle size.

A scanning electron microscopic photograph of a dehydrated stableemulsion that originally contained 70% moisture is shown in Fig. 8. In thismicrophotograph, the holes are the sites of oil droplets. Because theemulsion was dried at room temperature, the droplets did not change insize. The holes represent the actual size of the droplets. They show thatalmost all of the droplets have a diameter smaller than 1 mm.

For the beverage emulsion, the determination of particle size distribu-tion serves two purposes: One is to estimate the quality of the emulsionconcentrate and the other—even more important—is to predict the stabilityof the emulsion when it is made into the finished beverage at a future date.The latter is for quality control. It is to predict the stability of an emulsionby comparing its particle size data with the quality control standardestablished previously for this product.

There are many methods available for the determination of particlesize and size distribution. Groves and others (84,85) have reviewed thesemethods of particle characterization. Groves’ review emphasizes thosemethods that have practical use for emulsions containing smaller particles.

Figure 8 Scanning electron microphotograph of a dehydrated stable beverage

emulsion concentrate.


The beverage industry commonly uses the following methods for productdevelopment and quality control: optical microscopic method, transmit-tance measurement, and laser diffraction technique.

A. Optical Microscopy

The optical microscope is one of the most valuable tools for observing themicrostructure of emulsions (86). It is invaluable for a quick examination ofthe emulsion when there is no other particle size measurement instrumentavailable. Although the microscope is mainly used for examination of theoil particles, it is also useful for checking the gum solution. If there areinsoluble materials or underdissolved gum particles in the solution, theycan be seen under microscope. Because beverage emulsions are oil particlesdispersed in a transparent liquid and they have very similar refractiveindexes, a conventional bright-field optical microscope is not suitablefor examining oil-in-water emulsions. A microscope equipped with aphase-contrast attachment is best for this purpose. A microscope equippedwith a dark-field attachment can be used also. Both phase-contrast anddark-field microscopes will enhance the edge contrast of the image of oildroplets (87). When using the phase-contrast microscopy, an experiencedtechnician can view and estimate the particles down to 0.5 mm at 1000�magnifications with a micrometer. The theoretical limit of resolution of anoptical microscope is about 0.2 mm, but, in practice, it is difficult to obtainreliable measurements below about 1 mm. The experience of routine use ofan optical microscope is essential in particle characterization, even if it isused only to examine the state of dispersion for making crudemeasurements. In practice, it is common to use a microscope in conjunctionwith other instrumental analysis. Using an optical microscope for objectivemeasurement is almost impossible. A statistically significant count ofparticles, allowing a 5% error and at 95% confidence level, would need2960 particles. In general, the emulsion has to be diluted before beingplaced on the microscope slide. Brownian movement of particles usually isa problem for particle counting under the microscope. To slow down theBrownian movement, a gelatin solution, low-viscosity sodium carboxy-methylcellulose (CMC) solution, or glycerin in water are commonly usedto dilute the emulsion in place of pure water.

B. Light-Scattering Determination

In general, two types of light scattering method are used: (a) angular lightscattering, where the intensity of the scattered light is measured as a functionof the scattering angle, and (b) transmittance or turbidity measurement,


where the intensity of the transmitted light is measured. In principle, theangular light-scattering method contains more information, but turbiditymeasurement is easier to carry out using a spectrophotometer (88,89). Forroutine particle size analysis or stability tests, where precision is not highlycritical, turbidity provides a convenient monitor of the state of the emulsion.However, turbidity measurement only translates to average particle size.

In the soft drink industry, turbidity or transmittance measurement isused as a quick method to check the quality of the emulsion. Some othermethods are measuring the spectral absorption of diluted emulsion at twowavelengths, such as 850 nm and 450 nm, or 800 nm and 400 nm, and usingthe ratio of the readings as stability indices (90,91).

C. Laser Diffraction Technique

The particle counting technique has been known for many years in emulsionscience and pharmaceutical research (92). Providing adequate care insampling and preparing the electrolyte solution, particle size analysis canproduce accurate and reproducible results (84,92). In our study we foundthat there is good corelationship of the data of emulsion particle size dis-tribution obtained from the particle size analysis with the stability of theemulsion in beverage (65). It is generally agreed that the particle sizeanalyzer is capable of making a unique and valuable contribution to thesubject of particle size characterization (84,93). With the development inthe laser diffraction technique, several instrument manufacturers havedeveloped more sensitive and accurate particle size analyzers. Theseinstruments can determine particle size down to 0.01 mm with the uppersize limit to 1000 mm. It is important to have an instrument with theupper size limit at 1000 mm as to make it useful for determine the oil dropletsize in premix or crude emulsions. These instruments measure particlesize distributions by measuring the pattern of laser light scattered by theparticles in the sample. Arrays of photodetectors detect and measurethe intensity of the scattered light. Each particle’s scattering pattern ischaracteristic of its size. The pattern measured by the instrument is thesum of the pattern scattered by each constituent particle in the sample.When the duration of the measurement is long enough that the flux patternaccurately represents the contributions from all particles, an analysis of theresulting pattern will yield the particle size distribution of the sample.Because the measurement procedure is fully automated, once the emulsionsample has been placed in the instrument, it takes only a few minutes tocomplete. The data generated have proven to be invaluable for qualitycontrol and troubleshooting the processing of the emulsion.


For expressing the particle size distribution of an emulsion, it ispreferable to use volume percentage distribution rather than populationpercentage distribution. The reason is that volume percentage distinctivelyshows the quantity in volume of the large particles present in the emulsion.It can be clearly illustrated in the following example. In one emulsion, thereare 100,000 dispersed particles of 1 mm diameter and 100 particles of 10 mmdiameter. By percentage in population, 100, 10-mm-diameter particles areonly 0.1% in number of the total particle population. When expressed involume percentage, these 10-mm-diameter particles represent 50% of thetotal volume of all the particles. This is because the volume of 100,000,1-mm-diameter particles is equal to that of the 100, 10-mm diameter particles.In the beverage, if these 10-mm-diameter particles float to the surfaceand ring, it represents 0.1% of the particles population in the bottle. Thepercentage of 0.1% is almost insignificant in the ratio of the population.However, in reality, it represents 50% of the total oil volume and it presentsa serious ringing problem.

There is another way to express the particle size distribution data—that is, the total ‘‘surface area,’’ in square meter (m2), of the particles in eachsize category (93). When one oil particle is broken to many smaller particles,the total area of these particles will increase according to their reduced size.However, this method of data presentation is not as commonly used as thevolume percentage method. The reason is that it is not as easy and simple toassociate the data with the quality of the emulsion as the volume percentage.

VII. STABILITY CRITERIA SETTING

To set the criteria of particle size distribution for determining the stabilityof an emulsion, one has to realize that the criteria are different for eachemulsion because each emulsion has a different composition which influ-ences the physical and chemical properties of the product. Examples ofparticle size distributions of a typical stable emulsion and an unstableemulsion of orange oil with ester gum in gum arabic solution are shownin Figs. 9 and 10. In Fig. 9, the particle volume percentage of the mostcritical channel, channel 2 (0.4–0.5 mm), is over 50%, and the total volumepercentage of channels 1–5 (0.32–1.0 mm) is over 80%. In Fig. 10, theparticle volume percentage of channel 2 is slightly over 30%, and thetotal volume percentage of channels 1–5 is only over 60%, and there is aheavy-tail portion of particles larger than 1.0 mm. Figure 11 is the compositepicture of three histograms. The top one shows the particle size distributionin volume percent of an unstable emulsion. This emulsion was used to makea soft drink. On standing, this bottle of soft drink developed a ring in the


Figure 9 Particle size distribution histogram in volume percentage of a stable

emulsion of orange oil plus ester gum in gum arabic solution.

Figure 10 Particle size distribution histogram in volume percentage of an unstable

emulsion of orange oil plus ester gum in gum arabic solution.


neck of the bottle. The particle size distribution in volume percent ofthe creamlike ringing material in the neck region is shown in the middlehistogram. The bottom histogram shows the particle size distribution involume percent of the stable emulsion in the nonringing portion of thebottle. The middle and bottom histograms clearly illustrate the particle

Figure 11 Particle size distribution histogram of (top) an unstable emulsion;

(middle) the ringing cream section from a soft drink prepared from the above

emulsion; (below) the stable section of the soft drink in the same bottle.


size distributions difference of the stable portion and the unstable portion ofthe soft drink made of an unstable emulsion. It has been determined that theparticle size distribution shown in the bottom histogram can be used as thecriterion for stability for this particular emulsion. No matter which particlesize distribution criterion is used, it is first necessary to collect a series ofdata with regard to the relationship of different particle size distributionpatterns to their degrees of ringing. After these data have been obtainedfrom the soft drink system, then the stability criteria can be developed.

VIII. SUMMARY

Beverage emulsions are a unique class of emulsion. They are prepared in theconcentrated form first and then diluted in order to prepare the finishedbeverage. It has to be stable in both the concentrated and the diluted forms,as the water phases are of different compositions. The stability of the emul-sion can be achieved by the application of colloidal chemistry principles.

ACKNOWLEDGMENTS

The author thanks Ms. Joanna Wu Holmes for her excellent technicalassistance. The constructive comments on the preparation of this chapter byDr. Chi-Tang Ho are greatly appreciated.

REFERENCES

1. J. Holleran, New beverage trend: the changing face of fortified beverages,

Stagnito’s New Product Mag. July 28–30 (2000).

2. Anon, Beverage Industry, October, 14 (1987).

3. M. E. Carey, T. Asquith, R. S. T. Linforth, and A. J. Taylor, Modeling the

partition of volatile aroma compounds from a cloud emulsion, J. Agric. Food

Chem. 50, 1985 (2002).

4. H. D. Preston, Clouding agents and heading liquors, in Developments in

Soft Drink Technology, Volume 1 (L. F. Green, ed.), Applied Science, London,

1978, p. 87.

5. V. R. Kaufman and G. Garti, Effect of cloudy agents on the stability and opacity

of cloudy emulsions for soft drinks, J. Food Technol. 19, 225 (1984).

6. T. J. Coomes, Cloud stability in soft drinks, Proc. Inst. Food Sci. Technol. 4(1),

33 (1971).

7. A. Oppenheimer, Clouding agents for the food industry, Food Prod. Dev. 5(3), 90

(1971).


8. A. Oppenheimer, Clarifying the clouding agent picture, Food Prod. Dev. 8(5), 60

(1974).

9. U.S. Food and Drug Administration, Glycerol ester of wood rosin, Code of

Federal Regulations, Title 21, Section 172.735.

10. U.S. Food and Drug Administration, Soda water; identity; label state-

ment of optional ingredients, Code of Federal Regulations, Title 21, Section

165.175.

11. H. I. Enos, G. C. Harris, and G. W. Hedrick, Rosin and rosin derivatives, in

Encyclopedia of Chemical Technology, Volume 17, Interscience Encyclopedia,

New York, 1970, p. 475.

12. Anon., Hercules Ester Gum 8BG, Product Data No. 7043-4, Hercules, Inc.,

Wilmington, DE, (1999).

13. Anon., Hercules Ester Gum 8BG—Glycerol Ester of Wood Rosin, Hercules, Inc.,

Wilmington, DE, (1997).

14. C. K. Simon (to Dominion Products, Inc.), U.S. patent No. 3,353,961,

1967.

15. Polak’s Frutal Works, Inc., British patent No. 1,537,160, 1978.

16. World Health Organization, Food Additives Series 1974, No. 6,186, WHO,

Geneva, 1974.

17. U.K. Food Advisory Committee, Report on the Review of the Use of Additives to

Cloud Soft Drinks, London: U.K. Food Advisory Board, 1985.

18. R. H. Goins and H. E. Davis (to Eastman Kodak Co.), U.S. patent No.

3,096,324, 1963.

19. Anon., Sucrose Acetate Isobutyrate, Product Bulletin, Eastman Chemical

Products, Inc., 1983.

20. S. J. Byrd, SAIB—The oldest new ingredient for the beverage market, in 2001

Annual Meeting of the Institute of Food Technologists, 2001.

21. Bush Boake Allen, Ltd., U.K. patent No. 1 118 019, 1968.

22. W. J. Krasavage, G. D. Di Vincenzo, B. D. Astill, R. L. Roudabush, and C. J.

Terhaar, Biological effects of sucrose acetate isobutyrate in rodents and dogs,

J. Agric. Food. Chem. 21, 473 (1973).

23. B. G. Procter, P. Dussault, and C. I. Chappel, Biochemical effects of

sucrose acetate isobutyrate (SAIB) on the liver, Proc. Soc. Exp. Biol. Med.

142, 595 (1973).

24. J. C. Phillips, J. Kingsnorth, I. Rowland, S. D. Gangolli, and A. G. Lloyd,

Studies on the metabolism of sucrose acetate isobutyrate in the rat and in Man,

Food Cosmet. Toxicol. 14, 375 (1976).

25. U.S. Food and Drug Administration, Sucrose Acetate Isobutyrate. Code of

Federal Register, Title 21, Section 172.833 (1999).

26. Anon., Frequently Asked Questions about SAIB Food-Grade, Kingsport, TN:

Eastman Chemical, 2000.

27. Anon., Resinogum DD, Product Bulletin Q 19, Rouen: Iranex France S.A.,

1979.

28. J. S. Mills and A. E. A. Werner, The chemistry of dammar resin, J. Chem. Soc.

51, 3132 (1955).


29. J. S. Mills, The constitution of the neutral tetracyclic triterpenes of dammar

resin, J. Chem. Soc. 52, 2196 (1956).

30. Hasegawa Co., Japanese patent No. 4 834 233, 1973.

31. G. de Haas, Dutch patent No. 65,604, 1950.

32. R. L. Jolly, French patent No. 2 095 093, 1972.

33. N. H. Mermelstein, Water purification for beverage processing, Food Technol.

26, 47 (1972).

34. Anon., Production and Quality Control Manual, Dr. Pepper Co., Dallas,

Texas, in Beverages: Carbonated and Non-carbonated (J. G. Woodroof and

G. F. Phillips, eds.), AVI, Westport, CT, 1981.

35. C. D. Morelli, Product water–beverage, in Water Manual, 3rd ed. Beverage

World Publication, Great Neck, NY, 1994, Chap. 5.

36. N. Garti, Hydrocolloids as emulsifying agents for oil-in-water emulsions,

J. Dispers. Sci. Technol. 20(1&2), 327 (1999).

37. D. M. W. Anderson, Water-soluble plant gum exudates—Part 1: Gum arabic,

Process Biochem. 12, 24 (1977).

38. W. Balke, Types of gum acacia and their uses, in Meeting of the Society of Soft

Drink Technologists, 1979.

39. K. L. Krumel and N. Sarkar, Flow properties of gums useful to the food

industry, Food Technol. 29(4), 36 (1975).

40. M. Glicksman, Present and future uses of hydrocolloids in foodstuffs, in Gums

and Hydrosoluble Natural Vegetal Colloids, Fourth International Symposium,

Iranex, Paris, 1976.

41. O. E. Araujo, Effect of certain preservatives on the aging characteristics of

acacia, J. Pharm. Sci., 21, 535 (1966).

42. A. H. C. Chun, R. S. Joslin, and A. N. Martin, The high hydrophile–lipophile

balance of gums, Part I, Drug Cosmet. Ind. 82, 164 (1958).

43. W. L. Guess, Determination of the hydrophile–lipophile balance of acacia using

various surfactants, J. Pharm. Sci. 50, 238 (1961).

44. E. Dickinson, V. B. Galazka, and D. M. W. Anderson, Emulsifying behavior

of gum arabic. Part 2: Effect of the gum molecular weight on the emulsion

droplet-size distribution, Carbohydr. Polym. 14, 373 (1991).

45. E. Dickinson, V. B. Galazka, and D. M. W. Anderson, Emulsifying behavior

of gum arabic. Part 2: Effect of the gum molecular weight on the emulsion

droplet-size distribution, Carbohydr. Polym. 14, 385 (1991).

46. A. K. Ray, P. B. Bird, G. A. Iacobucci, and B. C. Clark, Functionality of gum

arabic. Fractionation, characterization and evaluation of gum fraction in citrus

oil emulsions and model beverages, Food Hydrocolloids, 9(2), 122 (1995).

47. C. G. Caldwell and O. B. Wurzburg (to National Starch and Chemical Co),

U.S. patent No. 2,661,349, 1953.

48. C. Richard and C. Bauer (to Anheuser Busch, Inc.), U.S. patent No. 4,035,235,

1977.

49. P. C. Trubiano, Succinate and substituted succinate derivatives of starch,

in Modified Starches: Properties and Uses (O. B. Wurzburg, ed.), CRC Press,

Boca Raton, FL, 1986.


50. C. Chiu, (to National Starch and Chemical Co..), U.S. patent No. 5,185,176,

1993.

51. T. Otera, H. Tsubomoto, and P. C. Trubiano (to National Starch & Chemical

Co.), EC patent No. EP1149845, 2001.

52. D. M. W. Anderson, Gums and Stabilizers for the Food Industry (G.O. Philips,

D. J. Wedlock and P. A. Williams, eds.), Elsevier Applied Science, London,

1986, Vol. 3, p. 79.

53. K. A. Coia and K. R. Stouffer, Shelf life study of oil/water emulsions using

vairous commercial hydrocolloids, J. Food Sci. 52, 166 (1987).

54. F. Thevenet, Natural sources of gum arabic, in Conference Londres, 1980.

55. E. Dickinson, J. Ma, and M. J. W. Povey. Creaming of concentrated oil-in-

water emulsions containing xanthan, Food Hydrocolloids, 8(5), 481 (1994).

56. D. J. McClements, Comments on viscosity enhancement and depletion

flocculation by polysaccharides, Food Hydrocolloids 14, 173 (2000).

57. T. E. Furia, Sequestrants in food, in Handbook of Food Additives (T. E. Furia,

ed.), CRC Press, Boca Raton, FL, 1972, p. 271.

58. W. Chantrapornchai, F. M. Clydesdale, and D. J. McClements, Optical

properties of oil-in-water emulsion containing titanium dioxide particles,

Colloids Surfaces A: Physicochem. Eng. Aspects 166, 123 (2000).

59. M. E. Carlotti, M. Gallarate, M. Trotta, and V. Canova, On the use of

saccharose esters in the preparation of O/W emulsions with liquid crystals,

J. Dispers. Sci. Technol. 21(1), 31 (2000).

60. W. D. Pandolfe, Effect of premix condition, surfactant concentration, and oil level

on the formationofoil-in-water emulsions,J.Dispers.Sci.Technol. 16(7), 633 (1995).

61. Anon., Homogenization and emulsification, APV Gaulin Technical Bulletin.

No. 67, 1983.

62. W. D. Pandolfe, Effect of dispersed and continuous phase viscosity on droplet

size of emulsions generated by homogenization, J. Dispers. Sci. Technol. 2,

459 (1981).

63. W. D. Pandolfe and R. R. Kinney, Recent developments in the understanding

of homogenization parameters, in 1983 Summer National Meeting, 1983.

64. Anon., The effect of the second stage homogenizing valve, AVP Gaulin

Technical Bulletin No. 33, 1974.

65. C. T. Tan and J. W. Holmes, Stability of beverage flavor emulsions, Perf.

Flavor 13(1), 1 (1988).

66. R. Chanamai and D. J. McClements, Impact of weighting agents and sucrose on

gravitational separation of beverage emulsions, J. Agric Food Chem. 48, 5561 (2000).

67. T. F. Tadros and B. Vincent, Emulsion stability, in Encyclopedia of Emulsion

Technology, Volume 1, Basic Theory (P. Becher, ed.), Marcel Dekker,

New York, 1983.

68. E. Shotton and R. F. White, Stabilization of emulsions with gum acacia,

in Rheology of Emulsions (P. Sherman, ed.), Pergamon Press, Oxford, 1963.

69. E. Dickinson, D. J. Elverson, and B.S. Murray, On the film-forming and

emulsion-stabilizing properties of gum arabic: Dilution and flocculation

aspects, Food Hydrocolloids 3(2), 101 (1989).


70. C. T. Tan, Beverage flavor emulsion—A form of emulsion liquid membrane

microencapsulation, in Food Flavors: Formation, Analysis and Packaging

Influences, (E.T. Contis et al., eds.), Elsevier Science, Amsterdam, 1998, p. 29.

71. D. Silber and S. Mizrahi, Coalescence of orange oil droplets at oil–gum solution

interface and some relationships to emulsion stability, J. Food Sci. 40, 1174

(1975).

72. D. J. McClements, General features of polymeric steric stabilization, in Food

Emulsions—Principles, Practice, and Techniques, CRC Press, Boca Raton,

FL, 1999.

73. R. Chanamai and D. J. McClements, Comparison of gum arabic, modified

starch and whey protein isolate as emulsifiers: Influence of pH, CaCl2 and

temperature, J. Food Sci. 67(1), 120 (2002).

74. J. Walford, Stabilisers for essential oils in flavour formations, Food Manuf.

51(2), 35 (1976).

75. V. R. Kaufman and N. Garti, Spectral absorption measurements for

determination of ease of formation and stability of oil in water emulsions,


76. F. Thevenet, Acacia gums—Stabilizers for flavor encapsulation, in Flavor

Encapsulation (S. Risch and G. Reineccius, eds.), ACS Symposium Series 370,

American Chemical Society, Washington, D.C, 1988.

77. P. Becker, Emulsions: Theory and Practice, Robert E. Krieger, New York,

1977, p. 119.

78. E. Dickinson and G. Stainsby, Colloid in Food, Applied Science, London,

1982, p. 44.

79. M. L. Layme, D. E. Dunstan, and M. L. Gee, Zeta potentials of gum arabic

stabilized oil-in-water emulsions, Food Hydrocolloids 13, 459 (1999).

80. S. C. Charma, Gums and hydrocolloids in oil-in-water emulsions, Food

Technol. 35(1), 59 (1981).

81. T. A. Iranloye, N. Pilpel, and M. J. Groves, Some factors affecting the droplet

size and charge of dilute oil-I-water emulsions prepared by self-emulsification,

J. Dispers. Sci. Technol. 4, 109 (1983).

82. S. Usui, Y. Imamura, and E. Barouch, Destabilization of oil-in-water emulsion

with inorganic electrolytes in the absence and the presence of sodium dodecyl

sulfate, J. Dispers. Sci. Technol. 8(4), 359 (1987).

83. T. M. Riddick, Control of Colloid Stability Through Zeta-Potential, Zeta-Meter,

Inc., New York, 1968, p. 22.

84. M. J. Groves, Particle size characterization in dispersions, J. Dispers. Sci.

Technol. 1(1), 97 (1980).

85. M. Kalab, P. Allan-Wojtas, and S. S. Miller, Microscopy and other imaging

techniques in food structure analysis, Trends Food Sci. Technol., 6, 180 (1995).

86. R. J. Mikula, Emulsion characterization, in Emulsions: Fundamentals and

Applications in the Petroleum Industry (L. L. Schramm, ed.), American

Chemical Society, Washington, DC, 1992, Chap. 3.

87. J. M. Aguilera and D. W. Stanley,Microstructural Principles of Food Processing

and Engineering, Elsvier, Amsterdam, 1990.


88. P. Walstra, Estimating globule size distribution of oil-in-water emulsion by

spectrophotometry, J. Colloid Interf. Sci. 27, 493 (1968).

89. W. D. Pandolfe and S. F. Masucci, An instrument for rapid spectroturbimetric

analysis, Am. Lab. 40, (1984).

90. K. Horie, S. Tanaka, and T. Akabori, Determination of emulsion stability by

spectral absorption I: Relationship between surfactant type, concentration, and

stability index, Cosmet. Toilet. 93, 53 (1978).

91. V. R. Kauffman and N. Garti, Spectral absorption measurements for

determination of ease of formation and stability of oil-in-water emulsion,


92. P. J. Lloyd, J. I. T. Stenhouse, and R. E. Buxton, Comparison of the errors

due to sampling and analysis by Coulter Counter, in Particle Size Analysis

(M. J. Groves, ed.), Heyden & Son, London, 1978.

93. J. K. Johnson, Droplet size analysis of soft drink emulsions by Coulter Counter,

in Proceedings of 29th Annual Meeting of the Society of Soft Drink

Technologists. 1982, p. 121.

94. B. Warburton, The Rheology and Physical Chemistry of Some Acacia Systems,

Monograph No. 24, Society of the Chemical Industry, 1966.


13Dressings and Sauces

Larry D. FordKraft Foods, Memphis, Tennessee, U.S.A.

Raju P. BorwankarKraft Foods, East Hanover, New Jersey, U.S.A.

David Pechak and Bill SchwimmerKraft Foods, Glenview, Illinois, U.S.A.

I. INTRODUCTION

Dressings and sauces include mayonnaise, spoonable and pourable saladdressings, condiment sauces (e.g., ketchup, barbecue sauce, spaghettisauce). The Association for Dressings and Sauces (ADS) reports the U.S.sales of dressings (including mayonnaise) in 2000 was $2.9 billion (U.S.)with a growth rate of 3.9%.

Food dressings vary widely in their composition, texture, and flavor.A listing of some of the most widely known dressings and their fatcompositions is given in Table 1. Dressings cover a broad spectrum ofoil–water composition and some products are defined on the basis of theiroil content. In the United States, standards of identity require that mayon-naise contain at least 65% vegetable oil by weight (some brands of mayon-naise contain 80% or more oil), a minimum of 2.5% acetic acid (vinegarfor microbial preservation; citric and malic acids may also be used pro-vided they are not greater than 25% of the weight of acetic acid), and egg-yolk-containing ingredients which may be liquid, frozen, or dried (the yolkprovides emulsifying properties and, in addition, gives the mayonnaise apale yellow color). Spoonable salad dressings may be very similar to


mayonnaise, but the standard of identity for these products requires aminimum of 30% vegetable oil and allows the use of starches as athickening agent (products generally contain 35–50% oil). Pourablesalad dressings, such as French dressing, may contain less oil and maycontain gums (1).

The Nutrition Labeling and Educational Act of 1990 (NLEA) allowsfor nutrient content claims on the food labels in the United States. The U.S.regulations for foods are published each year in Code of Food Regulations21 CFR Part 101. The three claims of particular interest here are Reduced,Low, and Free as applied to calories, sodium, fat, saturated fat, andcholesterol. Additionally, the NLEA standardized the serving sizes foreach product. Table 2 provides the definitions of these claims. Standardsvary by country.

Some fat-free salad dressings contain no oil and are not even emul-sions. For the purpose of this chapter, all ‘‘full-fat’’ dressings and saucesare considered as ‘‘classic’’ emulsions and their corresponding reduced-fatand fat-free counterparts are considered as ‘‘nonclassic’’ emulsions. Therationale behind this classification as opposed to the one based on fatlevel is that all so-called nonclassic emulsion-based products need to beformulated and processed to approach in attributes to their correspondingclassic emulsion-based counterparts. This need for matching the attri-butes presents its own unique and difficult challenges. In other words,challenges in making a 6% fat Ranch dressing are quite different than thechallenges in making a 6% fat mayonnaise. This chapter includes aspecial section on challenges encountered with nonclassic emulsionbased dressings and sauces.

Table 1 Typical Fat Contents of Dressings

and Sauces

Sample Percentage

Mayonnaise 75–84

Italian 50–60

Salad dressing (spoonable) 30–60

Blue cheese 30–40

French 36–40

Russian 30–40

Thousand Island 30–45

Italian (low calorie) 0–3

Barbecue sauce 1–2

Ketchup 0.1–0.2


Table 2 Nutrition Labeling and Educational Act Claims

Claim Definition Example

Reduced A nutritionally altered product contains at least 25% less of

a nutrient or 25% fewer calories than a reference food

Reduced calorie (25% fewer calories than reference

food)

Reduced sodium (25% less sodium than reference food)

Low A reference amount (and 50 g of food if reference amount is

small) contains 40 cal, 140mg of sodium, 3 g of fat,

1 g of saturated fat, and 15% of calories from

saturated fat, or 20mg of cholesterol

Low calorie (40 cal per reference amount of per 50 g, if

the reference amount is 30 g or 2 tablespoons,

whichever is greater)

Low saturated fat (1 g of saturated fat per serving or

50 g, whichever is greater, and 15% calories from

saturated fat)

Free A serving and the reference amount contains no or

physiologically inconsequential amount: <5 cal, 5mg

of sodium, <0.5 g of fat, <0.5 g of saturated fat, and

<0.5 g trans fatty acids, <2mg of cholesterol, or <0.5 g

of sugars

Fat free (<0.5 g per serving)

Sodium free (<5mg of sodium)

Sugar free (<0.5 g of sugars)


A. Microstructure of Dressings and Sauces

There is limited literature on the microstructure of sauces and dressings.Chang et al. (2) used, transmission electron microscopy (TEM) to examinethe interfacial film surrounding emulsified lipid droplets in diluted samplesof mayonnaise. They concluded that the film or membrane is composed ofcoalesced low-density lipoprotein of egg yolks and microparticles of eggyolk granules. They postulated that the stability of the lipid dropletsis attributed to the high degree of plasticity of particles and to fibrousmembranes on droplet surfaces.

Tanaka and Fukuda (3) demonstrated using scanning election micros-copy (SEM) that the addition of xanthan gum to French salad dressinginhibited lipid droplet fusion, which extended the shelf life up to 6 months.Tung and Jones (4), using light and electron microscopy, determinedparticle size (lipid droplet) in mayonnaise and a spoonable, starch-basedsalad dressing. The interfacial film on lipid droplets of diluted mayonnaisewas described. The morphology of the continuous phase could not bedetermined, however, because of the preparation techniques that wereused. Using SEM, they were able to describe the nonlipid material foundbetween droplets of a spoonable salad dressing (Fig. 1). Flukiger (5) studiedinternal phase volume dependence in mayonnaise using light microscopy. Itwas shown that clumping of lipid was dependent on the concentration ofemulsifiers.

Darling and Birkett (6) studied the role of fat crystallization in thereduction of emulsion stability in oil-in-water emulsions. Using freeze-frac-ture TEM, they showed the mechanism by which fat crystals penetrate theinterfacial membrane resulting in the breakdown of the emulsion.

The application of confocal scanning laser microscopy (CSLM) hasenabled Heertje et al. (7) to perform optical sectioning of mayonnaise. Thisinstrument allows a disturbance-free observation of internal structure withrelatively minor preparative steps for the sample. [Nine parts of a mayon-naise sample were mixed with one part of a Nile Blue solution (0.1%) andthe stained mayonnaise was placed between two glass slides for observa-tion.] The lipid droplets within the sample were tightly packed together,causing the outline of the droplets to be hexagonal in shape. Although thecontinuous phase was easily distinguishable, its structural components couldnot be determined because of relatively low resolution caused by magnifica-tion limitations.

Through a slight modification of an agar microencapsulation techni-que developed by Salyaev (8) (see also Refs. 9–11), high-resolution TEMs ofliquid and visocous samples in their near-natural or native states can beachieved.


B. Rheology of Dressings and Sauces

The discussion in this chapter will concentrate on applications to dressingsand sauces and will not address the basic concepts which have been dis-cussed elsewhere (see, e.g., Chapter 4 in F. 12, and Ref. 13). The techniquesused for rheological evaluation of dressings and sauces have depended onthe nature of the particular products.

Most dressings and sauces exhibit viscoelastic rheological behavior,although some ‘‘thinner’’ products are primarily viscous in their flow behav-ior. Mayonnaise and spoonable (or semisolid or viscous) salad dressings areexamples of products which show viscoelastic rheological behavior andalso possess a yield stress. Pourable salad dressings, such as French saladdressing varieties and Thousand Island dressing, are primarily viscous in

Figure 1 Salad dressing containing 49% oil viewed with scanning electron

microscopy. Lipid droplets (l) are embedded within a reticulum of amorphous and

fibrillarlike (arrows) material. The amorphous material is assumed to be cooked

starch paste. Scale bar equals 10 mm. (Courtesy of Dr. Marvin A. Tung.)


their flow behavior, but exhibit varying degrees of thixotropy and often ameasurable yield stress as well.

C. Stability of Dressings and Sauces

The emulsion stability of food dressings is a relative concept. All emulsionsin dressings are thermodynamically unstable and, given enough time, willseparate. Therefore, the formulator is fighting a losing battle and can neverdevelop a dressing with an indefinite shelf life. Indeed, the destabilizationmay even be involved in the release of flavor and the perception of mouth-feel. In addition, there are different types of emulsion stability. Someemulsions are formulated to give maximum stability against coalescence(mayonnaise) whereas other emulsions are formulated to give maximumstability against creaming (pourable salad dressings). Flocculation oraggregation, considered as instability in dispersion science, may actuallybe desirable in dressings and sauces.

In some emulsions, maximum stability is desired to keep productintegrity under adverse conditions (e.g., mayonnaise with an internalphase volume of 80%, which is past the point of hexagonal close packingof spheres at 74.05%). In other emulsions such as Italian dressing, onlyshort-term stability is necessary.

There are three main methods for stabilizing emulsions based on elec-trostatic, steric, and particle stabilization mechanisms. The theoretical treat-ments of these mechanisms are considered elsewhere (see Chapter X). In thischapter, a brief outline of the mechanisms and their relevance to saladdressings will be discussed.

D. Processing of Dressings and Sauces

Many of the dressings and sauces, being emulsions, involve an emulsifica-tion step. The energy imparted varies with the type of product—the desireddroplet size being of primary importance. Shear, turbulence, and cavitationalone or in combination are involved. Different types of emulsificationdevice are employed, colloid mill being the most common. This section isadded to this revised edition to discuss key factors controlling this criticalunit operations and present guidelines for the selection of emulsificationequipment.

E. Nonclassic Dressings and Sauces

Many of the dressings and sauces are high in fat and of nutritional con-cern to the consumers. As a result, the Better-For-You segment of these


products is growing. As the fat is reduced or eliminated and replacedby other nonlipid ingredients, the need to match the appearance, texture,mouth-feel, flavor, and performance (in recipes, if relevant) attributesof the nonclassic products with their classic counterparts presentsunique challenges. The prevalent thinking during the early developmentof fat-free products was that one must look for the magic replacementingredient that delivers all characteristics of fat. This was subsequentlyreplaced by a systems approach where specific technologies are leveragedfor different functions of fat to create a system that works in a syner-gistic fashion. The most critical remaining challenge to be overcome inthe nonclassic dressings and sauces is around flavor—deliver theright flavor impact and maintain its stability over the shelf life of theproduct.

II. MICROSTRUCTURE OF DRESSINGS AND SAUCES

A. Introduction

The tools and techniques used to investigate the microstructure of foods arevast and in some manner only limited to the imagination and creativity ofthe investigator. Relatively simple methods based on routine light micros-copy techniques have been demonstrated by Flint (14), where she separatesmethods based on what information is being sought in the investigations offood emulsions. If the ingredients present in a product are the main focus,the methods can allow for significant disruption of the product, selectedoptical setup (i.e., polarized optics) and selective stains (i.e., Oil Red O,toluidine blue, iodine) and result in a quick diagnosis. If, on the otherhand, the information being sought requires the product to remain intact,other methods are demonstrated using vapors such as iodine for starch andosmium tetroxide for oils. For these methods, the samples are prepared justprior to observation and the degree of success is in part based on the skilland experience of the investigator.

Other methods utilizing only slightly more sophisticated light micro-scopes employ the use of fluorescence optics. The sample preparationtimes can be as simple and quick as those for routine light microscopybut with the added specificity of molecular markers that can detectminute concentrations of minor emulsion components. A detailed descrip-tion of fluorescence microscopy as a technique applied to food systems canbe found in Ref. 15.

An example of the use of fluorescence microscopy being used to inves-tigate the proteins at the interface of emulsions can be found in the work ofSengupta and Damodaran, although this work is not specific to dressings


and sauces, the methodology could just as easily be adapted to these foodsystems. The authors used epifluorescence microscopy to demonstrate phaseseparation of mixed-protein films. By labeling each protein with a differentfluorescent compound, the authors were able to demonstrate that the initialsaturated monolayer exhibited a homogeneous mix of proteins but filmsthat were allowed to age for several days exhibited a redistribution of theproteins at an air–water interface.

To answer questions not addressed by the previously described tech-niques requires a significant increase in time and capital investment.Although the use of fixatives and embedding materials has allowed forthin sections of emulsions to be routinely achieved, the investment insample preparation time and tools to perform the methodology issignificant. Such methodology does, however, allow one to address newand possibly more challenging questions. The introduction of glutaralde-hyde and plastic embedding resins opens the door to fine-structure analysisthat can take advantage of transmission electron microscopy as well as lightmicroscopy observations.

Again, as with the light microscopy methods described earlier, thedetails of the methodology employed is greatly dependent on the questionsbeing asked. The use of agar to entrap mayonnaise prior to fixation andembedment was sufficient in the work of Tung and Jones (4), where the lipiddroplets and their associated interfacial film were the main focus of studyeven though the long-range microstructure was lost.

Other more elaborate techniques have been developed in an attemptto preserve long-range structure. One such method, developed by Salyaev(8), was designed for preparation of liquid, semiviscous, and viscous sam-ples. The critical element in this beautifully simple technique is the produc-tion of small, delicate, agar cylinders within which the sample is placed.The cylinders are then sealed at both ends and processed intact for exam-ination using standard methods. Samples processed in this manner remainin their natural or near-natural morphological state and the introductionof artifact is kept to a minimum. The best example of well-preservedmorphology is with mayonnaise. Using this method, the tightly packedlipid droplets retain their hexagonal morphology when viewed in crosssection (Fig. 2a).

Confocal laser scanning microscopy (CLSM) has the advantage ofallowing multifocal plane imaging of intact, often unprocessed samples.Although used extensively in the biomedical realm, its use in foodscience has been more slowly embraced. Heertje et al. (7) showed itspotential in food systems, and, later, they (17) demonstrated its useful-ness in studying the dynamics of emulsifier displacement at an oil–waterinterface. More recently and as an example of how microscopy data can


benefit from incroporation in a multifaceted approach utilizing imageanalysis and rheological and sensory data, Langton et al. (18) studiedtexture variations in mayonnaise. They varied the homogenization andtemperature levels, as well as the amount of egg yolk to produce prod-ucts that had defined organoleptic differences and then analyzed themusing CLSM and freeze-fracture TEM methods. Whereas, CLSM utilizedintact wet samples and the freeze-fracture TEM method utilized a quick-freeze method, they both avoid the more typical methods that disruptthe sample or allow for extensive repositioning of mayonnaise compo-nents. The authors were able to correlate fat droplet size and egg yolkdistribution with storage modulus and perceived texture.

The most common morphological feature of dressings and sauces isthe lipid droplet. Although there is great variation in the size of thedroplets (Table 3), all are coated with an interfacial film. When viewedin profile at high magnification with an electron microscope, the interfacialfilm appears as a thin electron-dense band. The width of the film rangesfrom 100 to 200 A depending on the type of sauce or dressing. An aqueousphase surrounds the lipid droplets. This phase is continuous and,

Figure 2 The microstructure of mayonnaise containing 80% oil. Low-magnifica-

tion transmission electron micrograph showing the tight packing of lipid droplets (l)

within sample. Note electron dense material between droplets. Scale bar equals 2 mm.

(Courtesy of R. S. Unger.)


in dressings, contains the spices and plant material that enhance flavor ofthe product.

Polymorphic masses of starch are present in the continuous phase ofsome spoonable salad dressings (Fig. 3). In mayonnaise, fragments of eggyolk granules are the predominant structure in the continuous phase. Thesefragments or particles adhere to the interfacial film and to each other,resulting in the formation of a protein network. The network increasesthe viscosity of the product and enhances stability of the emulsion. Saucessuch as ketchup and barbecue sauce contain a low percentage of lipid and ahigh percentage of water and plant material. In ketchup, the plant materialis presumed to be tomato cell walls. As a result of processing, the cellulosefibers that once comprised the cell walls become disassociated from eachother and form a fine network that can only be visualized with the electronmicroscope.

B. Mayonnaise and Spoonable Salad Dressings

Mayonnaise is an oil-in-water emulsion that is difficult to examine ultra-structurally because of its high lipid content (Table 1) and the fragility of itsinterfacial film. Special techniques (described earlier), must be employed toavoid the introduction of artifact caused by specimen preparation. Properlyprepared mayonnaise will contain lipid droplets that are tightly packedtogether and hexagonal in shape (Fig. 2). Lipid droplets in the sampleexamined have a mean diameter of 2.64 mm (2.0 SD, Table 3) and aresurrounded by an interfacial film that is approximately 140 A in width(Fig. 2). The continuous phase, located between lipid droplets, is composedof electron-dense particles in an aqueous environment. The particles are

Table 3 Lipid Droplet Size in Dressing and Sauces Determined by Light

Microscopy

Sample Sizea (mm) Variance

Blue cheese 10.8 6.06

Thousand Island 14.8 7.05

Russian 35.8 20.98

Italian (low calorie) 26.2 13.13

French 38.0 19.17

Italian 41.3 28.24

Salad dressing (spoonable) 1.96 1.37

Barbecue sauce 13.19 3.31

Mayonnaise 2.64 1.97

aSize measured as mean diameter.


Figure 3 The microstructure of starch-stabilized salad dressing. Low-magnifica-

tion transmission electron micrograph showing the wide size distribution of lipid

droplets (l) within the sample. Note aggregates of starch (s) with the continuous

phase. Scale bar equals 1.0 mm. (Courtesy of D. Dylewski and R. Martin.)


polymorphic in shape, average 550 A in size, and are presumed to be frag-ments of egg yolk granules. The protein particles adhere to one another,forming ‘‘bridges’’ between lipid droplets and also forming a layer or coat-ing on the interfacial film. The protein bridges undoubtedly influence therheological properties of mayonnaise, and the coating of the film shouldenhance emulsion stability.

The spoonable salad dressing examined in this study is an oil-in-wateremulsion that contains 46% lipid (Table 1). Some spoonable salad dressingshave a starch base. Lipid droplets in these samples are spherical to angularin outline and have a mean diameter of 1.9 mm (Table 3). The interfacialmembranes are continuous, approximately 120 A in width, and appear elec-tron dense. Polymorphic aggregates of starch are present in the continuousphase (Figure 3). At high magnification, small particles of starch adheringdirectly to the interfacial membrane can be resolved. The presence of starchin the continuous phase is thought to increase the stability of the emulsion.

C. Pourable Salad Dressings

Pourable salad dressings are emulsions that share one common morpholog-ical feature, the lipid droplet. Because lipid droplets frequently and rapidlycoalesce, depending on the type of dressing, morphological descriptions aredifficult from a technical standpoint. Phase separation further adds to thedifficulty and can only be overcome by using methods that are rapid.Differential interference light microscopy and freeze-fracture transmissionelectron microscopy are the methods of choice.

The lipid droplets are spherical in shape, surrounded by an interfacialfilm of various widths and composition, and vary in size (Table 3). Lipiddroplet size is specific for each type of dressing and is determined by thestabilizers used and the method of processing.

Based on lipid droplet size, the pourable salad dressings examined inthis study can be divided into two classes. Blue cheese, Ranch, andThousand Island are included in the first class of ‘‘creamy dressings’’ withdroplets that range from 10 to 15 mm in size or even lower. The secondclass contains droplets that are from 25 to 40 mm in size and includes‘‘oily dressings’’ such as Russian, Italian, and French dressings.

D. Sauces

1. Ketchup

The samples of commercial ketchup that were examined were composedprimarily of tomato cell fragments (Figs. 4a and 4b). Large fragments of


cell walls and aggregates of intercellular material all within the size range of10 mm or less could be resolved using light microscopy. Although theketchup was known to contain 0.1% oil, distinct droplets were not detectedin the sample.

Examination of thin sections using TEM revealed the presence of amajor structural/functional component of ketchup that was not resolvableat the light microscopic level. This component is a fine fibrillarlike materialthat is dispersed throughout the sample and presumed to be cellulose fibers

Figure 4 The microstructure of ketchup. (a) Light micrograph of a thick section

showing tomato cell fragments (arrows). Scale bar equals 10 mm. (b) Transmission

electron micrograph showing aggregates of plant cell fragments (arrows). Note: Cell

fragments appear to be embedded within a matrix of fine fibrillarlike material

(arrowheads). Scale bar equals 0.5 mm. (Courtesy of D. Dylewski and R. Martin.)


from disrupted plant cell walls. The fibers form a matrix within whichintracellular fragments are embedded. Occasionally, the fibers are aggre-gated with their long axes oriented in the same direction. We speculatethat these aggregates are partially disrupted cell walls. The fibers functionas water-binding agents, much like gums.

2. Barbecue Sauce

The barbecue sauce presented in this study is an emulsion that containsapproximately 1.5% oil (Table 1). The mean diameter of the lipid dropletsis 13.9 mm (Table 3), and they appear spherical in shape (Figs. 5a and b). Thecontinuous phase of the sauce is composed primarily of plant cell fragments,presumably cell walls of tomatoes.

III. RHEOLOGY OF DRESSINGS AND SAUCES

A. Introduction

Over a very long time, Heinz ketchup commercials have depicted the thick-ness of their ketchup as an important and differentiating quality attribute

Figure 5 The microstructure of barbecue sauce containing 1.5% oil. (a) Thick

section of sample viewed with bright field light microscopy. Lipid droplets (l) appear

to be entrained within plant cellular debris (arrows). Scale bar equals 10 mm. (b)

Differential interference micrograph showing fine dispersion of lipid droplets (l).

Scale bar equals 50 mm. (Courtesy of D. Dylewski and R. Martin.)


versus the competition. Barbecue sauces have, time and again, touted thick-ness as a good thing. Indeed, consumer acceptance of dressings and sauces isat least somewhat dependent on their texture. One very important reason tomeasure the rheology of dressings and sauces then is because the rheologicalproperties are related to the product texture.

Texture of dressings and sauces is very complex, multidimensionalterm and consumer liking of product texture is an overall effect stemmingfrom evaluations across all relevant textural attributes. In fact, as depictedin some Heinz ketchup commercials, the slowness with which the ketchuppours out of a bottle gives a visual assessment of the product thickness(textural attribute) and viscosity (rheological attribute). Thus, multiplesenses are involved in the assessment of texture. Hence, the need for manydifferent approaches and techniques to measure the rheological behavior ofthese products.

Even when considering a single attribute, and a simple one at that,such as thickness, the perception of that attribute is a combination of severaldifferent ways in which the attribute can be perceived. As discussed byBorwankar (19), consumer perception of the thickness of barbecue sauceis a combination of perception of viscosities from several different sensorycues: how the sauce pours out of the bottle, perception during basting, itscling, and, finally, its mouth-feel. Because the steady-state rheological beha-vior of the sauce is non-Newtonian, the viscosities relevant in these differentapplications are different because different shear rates are relevant (Fig. 6).This highlights the central difficulty encountered when one attempts todraw correlations between rheology and sensory perception: For a reason-able chance of success, only one means of sensory assessment of thetexture attribute should be emphasized, a feat not easily accomplished inpractice.

Perhaps the most emphasized viscosity characteristic for dressings andsauces is the viscosity relevant in the mouth. According to Shama andSherman (20), the relevant shear stresses and shear rates involved are depen-dent on the product’s viscosity itself, ranging from very high shear rates

Figure 6 Shear rates operating under different processes involved in sensory

perception of barbecue sauce.


corresponding to the shear stress of about 100 dyn/cm2 for very thin systemsto shear rate of about 10 s�1 for very thick systems. For typical pourabledressings and sauces, the shear rate range in question is about to s�1. Forspoonable dressings, firmness or thickness is perhaps more related to theyield stress rather than viscosity.

A more complex but often mentioned textural attributes, again espe-cially for spoonable salad dressings, is creaminess. Creaminess likely hasmore than just textural connotations. Even as a textural term, its sensorialassessment is complicated and not fully understood. Nevertheless, Kokini(21) suggested that creaminess is determined by thickness and smoothness,which he modeled via fluid mechanical models. He obtained a reasonablygood correlation between creaminess and rheological properties across sev-eral product categories.

B. Rheological Measurements

Viscosity is the characteristic most commonly used in salad dressingquality control (22) and the rheology of dressings and sauces has been thesubject of many articles. Although these have appeared in a number ofscientific and technical journals, the reader is referred especially to theJournal of Texture Studies and the Journal of Dispersion Science andTechnology.

The techniques used for rheological evaluation of dressings andsauces have depended on the nature of the particular product. For visco-elastic mayonnaise or spoonable salad dressings, various authors haveemployed creep compliance (see, for example, Refs. 23–28) and otherdynamic testing (see, for example, Refs. 29–32). Coaxial cylinder visco-meters and extrusion rheometers have been used to evaluate both the visco-elastic products and the more fluid dressings and sauces (see, forexample, Refs. 33–34).

Atkin and Sherman (23), from creep compliance studies, found thatmayonnaise exhibits nonlinear viscoelastic behavior at stresses of about10,000 dyn/cm2. Later, Kisseoglou and Sherman (25) concluded fromcreep compliance measurements at about 70 dyn/cm2 that mayonnaiseexhibited linear viscoelastic behavior. Gladwell et al. (38) studied thedependence of the rheological behavior of soy oil in water emulsionsupon oil concentration and Gladwell et al. (39) studied the dependence ofthe creep/recovery behavior of oil-in-water emulsions upon disperse phaseconcentration. In each of those studies, regions of linear viscoelasticbehavior and regions of nonlinear viscoelastic behavior were also found.This illustrates the importance of carefully specifying the measurementconditions. Similarly, Rao (40) noted that Boger and Tiu (41) observed


time-dependent shear thinning (thixotropic) behavior with mayonnaise,were Blake and Moran (33) did not. Rao (40) gave that example to demon-strate the necessity for determining whether a given sample has time-complicating factor in the measurement of viscosity of thick systems suchas mayonnaise is the occurrence of wall slip effects. The presence of thesehave been demonstrated by Pascual et al. (42), who recommend the use ofserrated configurations to minimize these and improve the reproducibility ofmeasurements.

Although Barnes and Walters (43) have argued convincingly that‘‘. . .given accurate measurements, no yield stress exists,’’ the yield stressconcept is useful in characterizing many dressing and sauces. As quotedby Barnes and Walters (43), Scott Blair (44) defined yield stress as ‘‘thatstress below which no flow can be observed under the conditions of experi-mentation.’’ Following this definition, various dressings and sauces do exhi-bit yield stresses when one considers the ‘‘conditions of experimentation’’ tobe the conditions of use by the consumer. Although many different ways areavailable for measuring the yield stress of a spoonable salad dressing, per-haps the simplest method is the vane method. This methid is best describedby Alderman et al. (45). Park, et al. (46) (see also Ref. 47) used a falling-needle viscometer and a rotating-concentric-cylinder viscometer to deter-mine the yield stress of several samples, including tomato ketchup. Thefalling-needle viscometer data led to a yield stress of 10.85N/m2. Perhapsthe best technique to measure the yield stress of pourable dressings such asRanch is using a controlled-stress rheometer.

Recently, a spate of new rheological methods is being applied to foodsystems, especially dressings. Accounts of extensional viscosity measure-ments and tribology (thin layer) measurements are appearing in the litera-ture (see Refs. 48 and 49, respectively). These are emerging areas and thesignificance of these rheological properties for food scientists and engineerswill become clearer as these techniques are more widely utilized.

C. Emulsion Rheology

For a good overview of emulsion rheology, refer to the work of Sherman(12, Chapt. 4; 50, 51). A general representation of viscosity of emulsions canbe written as

� ¼ f ð�0,�, d,AÞ ð1Þ

where � is the viscosity of emulsion, �0 is the continuous phase viscosity, � isthe phase volume fraction of the dispersed phase, d is the drop size, and A isthe state of aggregation of the emulsion. This equation is a very good way of


thinking about emulsion viscosity, although it clearly is a simplification. Theemulsion viscosity increases with the phase volume fraction of the dispersedphase in a non-linear manner. The phase volume fraction of the dispersedphase includes contribution from the thickness of the emulsifier layer whosecontribution increases as the drop size decreases.

The effects of drop size and aggregation are primarily seen with rela-tively concentrated emulsions ð� > 0:2�0:4Þ. The viscosity of the emulsiondecreases with increasing drop size. This is due to a number of reasons:hydrodynamic resistance dependent on the average separation distancebetween drops, contribution from interfacial rheology, electroviscouseffects, increased aggregation seen in finer emulsions, and contribution ofinterfacial layer thickness mentioned earlier.

Aggregation increases the emulsion viscosity. The interdroplet inter-action determines if the aggregates are relatively open or compact. Openaggregates entrain large amounts of continuous phase and hence havehigher viscosity than the compact ones. Open aggregates occur when thereis a strong attractive interaction between droplets, and compact aggregatesare formed when the attraction is weak (see Ref. 52). The aggregation stateof an emulsion is clearly dependent on the shear rate and, for thixotropic(time-dependent) system, on shear history. As the shear rate increases,aggregates are broken down to smaller and smaller sizes and the emulsionviscosity decreases. Hence, concentrated emulsions often show a shear thin-ning (non-Newtonian) behavior.

Thus, the shear thinning seen in dressings and sauces is attributable toprogressively breaking aggregates with shear. Gums and stabilizers can playa role as well (see Section IV.C). The time-dependent or thixotropic behav-ior arises if the kinetics of breaking and forming aggregates is slow. ChanMan Fong and De Kee (53) have modeled this behavior and shown that themodel explains transient rheological properties of mayonnaise fairly well.Amemiya and Shoemaker (54) have shown that, for model dispersions,thixotropy increases with increasing phase volume fraction and decreasingdrop size.

The viscosity of emulsion increases as the continuous phase viscosityincreases. Typically, gums and stabilizers are added to pourable dressingsand salad dressings to increase the continuous phase viscosity and, hence theviscosity of the emulsion. Gums and stabilizers have non-Newtonian rheol-ogy and they impart non-Newtonian character to the emulsion even whenthe amount of the dispersed phase is low.

Equation (1) can actually be generalized to any rheological property,not just viscosity. For yield stress and shear modulus of concentrated emul-sions, Princen (55) derived equations theoretically that directionally confirmthe drop size and phase volume fraction effects. In their model, which did


not account for aggregation, yield stress and shear modulus are a result ofpacking the emulsion beyond the close-packing limit. Aggregation can actu-ally produce finite yield stresses and shear moduli even below the close-packing limit. Similarly, gums and stabilizers if added can have yieldstress and shear modulus as well themselves.

The above discussion actually suggests that aggregation can be used toone’s advantage in formulating dressings and sauces. An aggregated systemwill have a higher viscosity than a nonaggregated system. Thus, one canformulate a dressing (or a sauce) with less oil and/or less gums and stabi-lizers by causing controlled aggregation. As discussed in Section IV and inother chapters in the book, aggregation leads to an increase in creamingand coalescence. The control lies in devising an aggregated system that isotherwise stable to creaming and coalescence.

IV. EMULSION STABILITY OF DRESSINGS AND SAUCES

A. Introduction

Dressings cover a broad spectrum of oil–water composition, as is seen inTable 1; from mayonnaise, a 65–84% oil-in-water emulsion, to fat freedressings, which may contain no oil. Most commercially important saucescontain little oil and, therefore, will not be considered in this section.

Dressings also cover a diverse range of products. For discussion pur-poses, dressings are divided into three categories: semisolid, pourable, andnonclassic. Each category has its own requirements for emulsion stability,but each product shares the need to maintain the emulsion integrity duringprocessing, packing, transportation, storage, consumer preparation, andconsumption. Of these dressings, only mayonnaise, salad dressing, andFrench have their own standard of identity. Mayonnaise must have atleast 65% fat and salad dressing and French dressing must have 30% and35% fat, respectively. The oil phase volumes of these and other dressings areevident from viewing Figs. 1–5.

As previously mentioned, the emulsion stability of food dressings is arelative term due to the fact that all emulsions in dressings are thermo-dynamically unstable and, given enough time, undergo phase separation.However, in a kinetic sense, the emulsions in dressings can be made stablethrough an acceptable shelf life and still maintain appearance, texture, andflavor that are desirable to the consumer. For individual dressings, emul-sion stability may be concerned with any or all aspects of flocculation,creaming, and coalescence. The half-life of dressing emulsions may rangefrom seconds for a product like separating Italian dressing to years formayonnaise.


B. Theoretical Considerations

As discussed in Chapters 1 and 2, the primary modes of destabilization ofemulsions involve creaming, flocculation, and coalescence. These processesoccur concurrently and tend to build upon each other. Coalescence is pro-moted in cream layers and in flocs. Coalescence and flocculation lead toincreased creaming rates. Differential creaming promotes flocculation.Although traditionally considered as a mode destabilization in dispersions,flocculation or aggregation is not necessarily bad in itself. Flocculation can,in fact, be beneficial if it does not lead to visual separation and/or coales-cence. This is because aggregation leads to increased rheological propertiesand, thus, can provide a cost-savings opportunity (see above).

Emulsions are typically stabilized against creaming by the use ofstabilizers such as gums and starches. The primary factor leading to stabilityis the increase in continuous phase viscosity, which can proportionallyreduce creaming velocities, and, at times, the presence of a yield stress,which can completely suppress creaming. Gums are known to actually pro-mote aggregation by either bridging or depletion mechanisms, depending onwhether the gum attaches to the interface or not. Nevertheless, the suppres-sion of creaming due to increased continuous phase rheological propertiesfar exceeds any increase caused by promotion of aggregation.

Stability toward coalescence in the emulsions derives from the use ofemulsifiers which are surface-active agents and/or particles. For coalescenceto occur, the drops must encounter each other and the thin film trappedbetween the drops plays a crucial role. The drainage and stability of theintervening thin film controls whether or not coalescence will occur in suchencounters. The film can be stabilized either by reduction or elimination ofthe driving force due to repulsive interactions between droplets that cancounteract the forces that are pushing the drops together or by slowingdown the drainage rates.

Kinetic stability against coalescence can be obtained via a decreasein the rate of drainage of the intervening thin films due to suppressedinterfacial mobility. This is achieved either due to persistence of inter-facial tension gradients that suppress mobility, the Gibbs–Marangonieffect, and/or the interfacial viscoelasticity [see, for example, the reviewsby Ivanov, (56), Ivanov and Dimitrov (54), and Wasan (58)]. Thesechanges reduce the rate of drainage of the thin film between emulsiondroplets, thus suppressing coalescence. Although there has been con-siderable amount of work done on drainage and subsequent ruptureof thin films, the timescales involved in these processes are so short(seconds to minutes in most reported studies) that the kinetic mechanismis likely not important in providing the shelf stability to dressings and


sauces.* Rather, one relies on changing the interaction between the drop-lets by enhancing repulsion. The primary means of stabilization areelectrostatic, steric, structuring and particle stabilization. Several ofthese mechanisms are commonly recognized and are also described inChapter 1 and are only briefly covered here. Perhaps the one exceptionis structuring, which will be described in Section IV.B.3. These stabiliza-tion mechanisms rely on altering the interactions between droplets torender emulsions metastable.

1. Electrostatic Stabilization

The attractive forces of the van der Waals interactions and the electrostaticrepulsion forces due to the diffuse electrical double layer are the basis forthe DLVO (Derjaguin, Landau, Verwey, and Overbeek) theory of colloidstability.

The electrostatic repulsive force is derived from the local accumulationof counterions at a charged surface, the concentration of these ions beingstrongly dependent on the ionic strength (a function of salt concentrationand valence of the ions) of the medium. The thickness of the diffuse layer ofcounterions around the charged particle surface is compressed by the saltand hydrogen ion concentrations normally found in dressings to such adegree as to make the repulsive force ineffective in stabilizing the emulsion.Even though complex ionic surfactants such as proteins are commonly used,their method of stabilization extends beyond that of electrical repulsiveforces and will be covered in the next section. Therefore, electrostatic sta-bilization plays only on indirect role in dressing emulsion stabilization by itseffect on protein structure.

2. Steric Stabilization

As two emulsion droplets approach closely, the adsorbed surfactant layersinteract. For macromolecules adsorbed to the surface of emulsion droplets,only a portion of the large molecule is at the surface. Much of hydratedstructure remains in the aqueous phase. Therefore, as two emulsion dropletswith adsorbed macromolecules approach each other, the chains of themacromolecules interact in at least two ways. First, the number of config-urations that the molecular chains can attain is reduced (along with their

*On the other hand, this mechanism is likely to be very important in the formative stages of the

emulsion in preventing recoalescence (see later). The studies of thin film drainage in the early

stages of emulsion formation are lacking due to the inherent difficulties involved. However,

recently, some of the nonequilibrium effects involved have been investigated.


entropy). Additionally, the hydrated portion of the chains has associatedwater molecules which, when forced from the chain, create a local osmoticgradient (increases the enthalpy) which tends to force the particles apart.

For small-molecule nonionic emulsifiers, the lack of hydrated struc-tures extending from the surface reduces the entropic and enthalpic forcesuntil they are often too weak to provide adequate stability. However, small-molecule nonionic emulsifiers such as the polysorbates are used in pourabledressings with good results. Here, stabilization may instead be caused by thestructured packing of micelles in the continuous phase between approachingdroplets.

3. Structuring

In Chapters 1 and 2, the formation of liquid crystals was considered as oneof the ways that emulsifier structures can stabilize emulsions. A relativelynew mechanism recognized over the last decade or so is the stratificationphenomena seen in various systems. The presence of micelles, particles, orsimilar discrete entities in the bulk can lead to stratification in the thin filmsseparating emulsion droplets. This phenomenon has been discovered anddeveloped byWasan, Ivanov, and co-workers (see, for example, Refs. 58–60).The ordered structure occurs in thin films and leads to a stabilizinginteraction. This has been shown to occur in food dispersions as well (61).

4. Particle Stabilization

Emulsions can also be stabilized by adsorption of particles at the dropletinterface (62). It is the balance of energies at the solid–liquid and liquid–liquid interfaces which determine the effectiveness of the particle stabiliza-tion. In other words, both the oil and the water phases should prefer contactwith the solid particle rather than with each other. Theoretically, the opti-mum stabilization occurs when the oil and water phases have equal prefer-ence for the particle and the contact angle at the droplet surface is 90�C. Forpractical purposes though, the contact angle should be between 60 and 70�Cto overcome instability which could result from perturbations at the inter-face. Contact angles can be changed by adding surfactants which adsorbpreferentially to one interface.

Particle stabilization is thought to be present in several dressingsystems. For example, in dressing systems, finely ground spice particlessuch as mustard have been attributed to enhancing emulsion stability.Also, in mayonnaise, small particles from the egg yolk have been attributedas providing the major stabilizing force.


C. Emulsifiers and Stabilizers

As discussed later in Section V, the formation of droplets in dressing emul-sions has little value if the droplets are not protected from coalescence byemulsifiers (the half-life of emulsions without emulsifiers is about 1 s).Typically, stabilization is divided into two processes: transient (at the timeof formation) and long term (over shelf life). Small-molecule surfactants seemto impart better transient stability (due to their ability to migrate along thedroplet surface), whereas large-molecule surfactants function best at long-term stability. However, surfactants are seldom additive in their effects onemulsion stability. In fact, when more than one surfactant is present, theycompete for the interface. Under equilibrium conditions, the surfactant withthe greatest ability to lower the interfacial tension is preferentially adsorbed,and if present in sufficient quantity to cover all available interfacial area, itcan effectively prevent adsorption of other surfactants. Small molecules oftenadversely affect the long-term storage because they preferentially adsorb tothe interface and inhibit absorption of the large-molecule surfactants, whichare actually better at providing long-term stability.

Three general types of surfactants are used in dressings: substitutedpolysaccharides, polyoxyethylene derivatives of sorbitan fatty acid esters(polysorbates), and proteins (Table 4). The substituted polysaccharidessuch as propylene glycol alginate are used in various salad dressings toprovide low levels of emulsion activity and stability. Low-molecular-weight surfactants such as the polyoxyethylene derivatives of sorbitanfatty—acid esters—the polysorbates [high HLB (hydrophile–lipophile bal-ance)] are used in pourable dressings, where greater emulsion stability isdesired. Proteins are present in many dressing emulsions and are an impor-tant class of food surfactant. The manner in which proteins absorb andrearrange at the droplet interface is critical to emulsion stability.

Table 4 Sources of Emulsifiers

Emulsifier type Ingredients

Protein Buttermilk, sour cream, skim milk, nonfat dry milk, whole

milk, sodium caseinate, whey, whole eggs (fresh, salted,

frozen), egg whites (fresh, frozen, dried), egg yolks (fresh,

salted, sugared, dried)

Phospholipids Egg yolks, whole milk, sour cream

Particle Mustard flour

Synthetic Polysorbate

Chemical modified Propylene glycol alginate


Proteins form a film around the surface of oil droplets to give stableoil-in-water emulsions. The interfacial activity of a protein proceeds throughthe following stages: (1) native protein molecule diffuses to the interface,(2) protein penetrates the interface, and (3) molecules rearrange to achieveminimum energy. Stage 1 of emulsification by proteins is a diffusion-depen-dent process; therefore, any variables such as temperature, shape of theprotein, and viscosity of the medium will affect this stage. In stage 2, theprotein molecule arrives at the oil–water interface and causes a reductionin interfacial tension as it establishes the interface. As a general rule,the interfacial tension should be lowered below 10 dyn/cm for effectiveemulsification to take place (63).

After the proteins establish the interface, a slow change in interfacialtension with time is observed. Graham (64) explains that this change is aconsequence of the molecular rearrangements within the protein film.Rearrangement is fast for flexible proteins, such as casein, and slow forrigid globular proteins, such as lysozyme. Interfacial films from globularmolecules have more residual structures, such as helices, therefore allowingmore cross-linking and chain entanglement. This greater extent of cross-linking and entanglement produces a greater resistance to shear and dila-tion, thus resulting in greater resistance to shear and dilation thus resultingin greater viscoelasticity and stability. Globular proteins tend to form morecohesive films than flexible proteins. Films containing globular proteins willbe more stable than films containing flexible molecules because lateral cohe-sion of the globular proteins will tend to heal defects in the films. Forprevention of coalescence in protein-stabilized emulsions, the pH shouldbe away from the isoelectric point and the interfacial layers need to beheavily hydrated and electrically charged. The role of the thickness of theprotein layer, however, is a point of controversy.

As already mentioned, mayonnaise has an extremely high dispersed oilphase volume ranging from around 75% to 82% in commercial samples.Due to this high dispersed volume, the droplets are mostly in contact witheach other and the spherical shape of the oil droplets has become deformeduntil the drops actually resemble the polyhedral figures seen more often infoams. Obviously, whatever forms the barrier between adjacent dropletsmust be mechanically strong, rugged, and present a high-energy barrier tocoalescence. There are two mechanisms left to impart stability to the emul-sion: steric and particle. For steric stability, macromolecular surfactantssuch as protein are needed. For particle stability, particles with a polarportion for water and a nonpolar portion for the oil (perhaps a complexof lipid/polar-lipid/protein) are required. Fortunately, a natural emulsifierexists which contains components for both methods of stabilization—theegg. Typically, the batch process for making mayonnaise involves starting


with an aqueous phase of egg yolks, vinegar, salt, sugar, and mustard grainsto which vegetable oil is added slowly until a seed emulsion of desiredviscosity is formed. Then, the remainder of the oil may be incorporatedrapidly.

Gums and/or starches (see Table 5) are added to most dressings for avariety of reasons. These polysaccharides are not surface active (with theexception of PGA) but act to thicken the dressing which affects mouth-feel,cling, and so for (refer to Section III. C). Most importantly, stabilizers canimprove the emulsion stability of the dressing. Of course, by making thecontinuous phase more viscous, oil droplets encounter each other less often.More importantly, it is the creation of a yield stress which increases emul-sion stability by minimizing or eliminating creaming of the oil dropletsunder the relatively low force of gravity (65).

D. Natural Ingredient Effects

1. Eggs

There have been numerous studies on eggs and egg yolks, with regard totheir composition as well as the emulsifying properties. For an excellentreview, the readers are referred to Refs. 66 and 67. Egg yolks are themost functional components of the whole egg as far as the emulsifyingfunctionality in dressings and sauces is concerned. Egg yolk owes its emul-sifying activity to a lecithin protein complex (lipoproteins). Egg whites(albumen) are less functional. Egg yolk is a suspension of particles in protein

Table 5 Polysaccharides and Their Use in Dressings

Gums/starches Comments

Xanthan Most widely used gum; salt, acid, and heat resistant;

suspending agent; stabilizer; gelling reaction with

locust bean gum and guar gum

Sodium alginate Gels with Ca2þ ions; high viscosity with heat

Propylene glycol alginate Stabilizer; emulsifier; thickener; some gelling with

Ca2þ ions; pH tolerant;

Locust bean gum Thickener; gels with xanthan; insensitive to Ca2þ ions

Guar gum Thickener; insensitive to Ca2þ ions; cost-effective

Gum arabic Some emulsifier activity; thickener; stabilizer

Gum acacia Suspending agent; forms films at interface

Starch Thickener; retrogradation a problem

Modified starch Inhibits retrogradation; thickener

Microcrystalline cellulose Adds body and mouth-feel


solution. Although liquid egg yolk has been a standard raw material fordecades, many suppliers today are promoting egg yolk powders with excel-lent flavor and functionality. Today’s egg yolk powders owe this to a gentlespray-drying process used that prevents heat abuse. The egg yolk powdershave longer shelf life and do not deteriorate in performance and taste whenstored correctly. These can be readily hydrated prior to use.

Egg yolk contain components which contribute to emulsion stabilityby both steric and particle mechanisms. The particle mechanism is evident inmayonnaise in the micrograph shown in Fig. 2, where protein particles arefound at the interfaces between oil droplets. The isoelectric point of eggwhites and egg yolks are 5.4 and 5.3, respectively (68). At neutral pHthese egg fractions are negatively charged but at the pH’s of most saladdressings (pH 2.8–pH 4.0), they are positively charged. However, dueto the rather high ionic strengths of dressings due to high salt content,electrostatic stabilization is not likely.

2. Mustard Flour

Salad dressings typically contain mustard flour. Chang et al. (2) note that,according to Corran (69), powdered mustard is an effective emulsifier.

Fischbach and Kokini (24) examined the effects of variable mustardflour levels on oil-in-water emulsion stability and rheology. They found thatadding lower levels, (up to 0.5%) of mustard flour increased stability, buthigher levels of mustard flour (0.75% and 1.0%) led to decreased emulsionstability, possibly due to formation of a xanthan gum–mustard proteincomplex. Creep parameters also reached a maximum at lower levels ofmustard flour, depending on the age of the emulsion. Mustard flour isadded to products for its flavor contribution, but it is also thought to con-tribute to emulsion stability, possibly by a particle mechanism.

3. Dairy Proteins

Today, more and more salad dressings contain dairy ingredients incorpo-rated in them. Dairy proteins are mainly of two kinds: caseins and wheyproteins. Depending on the dairy ingredient used, whey proteins may or maynot be included. If cheese is used, proteolytic fractions of caseins may beincluded. All of the caseins and a whey protein, a-lactalbumin, have anisoelectric point in the range of 4.1–4.5 (70). The other major whey protein,b-lactoglobulin, has an isoelectric point of 5.3. As a result, the environmentof many dressings is very close to the isoelectric point of these proteins andthese proteins may be nonfunctional as emulsifiers in dressings. Worse yet,they may compete with the functional emulsifiers such as egg yolk for theinterface and actually cause emulsion stability issues.


E. Practical Considerations

As previously mentioned, in order to more easily understand the character-istics and requirements of emulsion stability for food dressings, it is worth-while to consider dressings as three categories: semisolid, pourable, andreduced-calorie dressings.

Semisolid dressings are, by definition, very viscous emulsions (refer tosection III. C). In mayonnaise, this is due to a high phase volume of oil,whereas in salad dressing, the effect of phase volume is augmented throughthe use of stabilizers (starches) which thicken the continuous phase. Due tothe high oil content of the mayonnaise emulsion, the oil droplets are forcedclose together, and in extreme cases, the spherical shape of the dropletsbecomes deformed (see Fig. 2). This close proximity precludes concern forflocculation, and emulsion stability must instead be concerned with prevent-ing coalescence. Because in may cases droplets are actually in contact withneighboring drops, coalescence can only be delayed by employing tough,pliable membranes around the oil droplets. In semisolid dressings, this mem-brane is formed from protein and polar lipid components of egg yolk.

Pourable dressings are, of course, much less viscous than semisoliddressings and as such are susceptible to creaming. The oil phase volumenormally ranges from 30% to 50% oil, but can be as high as 60–65% for thecreamy style dressings. Due to this lower ratio of oil to water, the oil dro-plets are spaced apart from each other and flocculation, which could lead tocoalescence, is also of concern.

Low-fat and fat-free dressings are generally based on the full-fat dres-sings with the addition of starches and gums to replace the oil. Of course, asthe oil is reduced, the emulsion stability does not become a problem; how-ever, maintaining the low amounts of oil in a fully dispersed state through-out the dressing can be challenging in lower-viscosity pourable dressingtypes. Special considerations have to be given to imparting a full-fatflavor and texture for these products as well. These are considered inmore detail in Section VI.

A prediction of the food emulsion’s stability without actually goingthrough shelf life is very desirable. Typical measurements have evolvedaround accelerated aging tests but must be viewed carefully because manyforces caused by the ‘‘aging’’ test would have never been experienced by anormally aged emulsion. One must realize that the accelerated physicalstability tests do not consistently accelerate many processes occurringthrough normal aging (e.g., oxidation and hydrolysis), which are just asimportant in determining product shelf life. Nevertheless, the assessmentof emulsion stability is often necessary. A typical protocol for assessingproduct’s emulsion stability should include several methods: shelf-life


testing at normal storage temperature, abuse testing at elevated and/orcolder temperatures than normal to simulate warehouse and consumerabuse and abuse testing vibration to simulate distribution abuse; perhapseven freeze–thaw cycles. Accelerated testing such as the centrifugal test maybe used as well, but with the above caution. The science behind the varioustests is not completely understood, making them an art rather than trulyscientific, but, often times, practitioners have developed enough productexperience that well-designed abuse tests can serve them well.

There is a real need for a more scientific prediction of emulsionstability, especially as new products and cost-reduced products are devel-oped at a frantic pace. Common methods based on the partitioning behaviorof surfactants such as HLB and PIT (phase inversion temperature) are oflittle use in most dressing systems. This is because the major stabilizingforces are usually due to proteins and adsorption of small-molecule surfac-tants interferes with their adsorption. Currently, of the four most promisingareas of predicting emulsion stability, two are based on properties of theinterface between the oil and water. For years, the correlation of interfacialrheology and emulsion stability has been claimed, but as discussed, thesuppression of the drainage rate that the interfacial rheology causes is notlarge enough to provide long shelf stability. More recently, Darling andBirkett (6) have given acclaim to the thickness of the interfacial layer as apredictor of emulsion stability. The new area of structuring in the thin film islikely to emerge as the most definitive predictor of stability. The final pre-dictor of emulsion stability is based on particle size and distribution andhow these change over time. It is actually an old method, but due to recentrefinements in resolution, it allows the researcher to follow coalescence in itsearliest stages.

The challenge to the food emulsion scientist is to comprehend the foodsystem on its many levels—not only those of emulsion science (molecularstructure of surfactants; assemblies and aggregates at the interface and inthe bulk; rheology; and stability) but their consequential effect on the orga-noleptic properties of the food as well (correlated with sensory evaluation).

V. PROCESSING OF DRESSINGS AND SAUCES

The primary goal of emulsion processing for dressings and sauces is toproduce a uniform, physically stable product with desired textural attri-butes. Although emulsion processing often has its greatest effect on producttexture and stability, important changes in product flavor, color, and sheenmay also occur. In general, reducing the mean internal-phase particle sizeand narrowing the particle size distribution will increase product stability,


viscosity, and yield. However, smaller oil droplets also refract light differ-ently than larger droplets and will result in a product that is lighter in colorand greater in opacity. Additionally, smaller oil droplets may release flavorsdifferently than larger oil droplets, possibly reducing the intensity and delay-ing the impact of flavor as the product is tasted. Therefore, the benefits ofproducing a more stable product by reducing droplet size must be weightedagainst the impact of texture, flavor, and color.

As discussed by Walstra (71), breakage of the larger droplets intosmaller ones is not the only process occurring during emulsification. Manyother processes occur simultaneously. Adsorption of the emulsifier on tonewly formed droplet surface is a key process that needs to occur during thetimescale of emulsification, otherwise the newly formed droplets that are notadequately stabilized by the emulsifier will recoalesce and all the disruptionis for a naught. In order to adequately stabilize the newly formed emulsiondroplets against recoalescence, first and foremost the emulsifier must bepresent in sufficient quantity to provide minimally monolayer coverage onthe interface and, second, the monolayer coverage must be reached duringthe timescale of the emulsification process. If the adsorption is too slow,recoalescence cannot be prevented.

Thus, as described in Fig. 7, drop breakage, emulsifier adsorption, andrecoalescence of the droplets are the three processes of importance. The finaldroplet size of the emulsion is governed by the balance of the breakage andthe recoalescence processes. The droplet breakage is primarily governed bythe time, amount, and distribution of energy input into the system. Therecoalescence step is primarily impacted by the type, amount, and adsorp-tion kinetics of the emulsifier. Efforts to optimize emulsification mustaccount for the interdependencies between the functional properties of theemulsifier and the processing equipment.

Figure 7 Processes occurring during emulsification. (Adapted from Ref. 71.)


The main steps in an emulsification process are (1) preparation of theoil and aqueous phases with proper incorporation of emulsifiers, hydrocol-loids, and other dry ingredients, (2) mixing the phases to form a uniformpremix, which may be in the form of either a coarse emulsion or an unstabledispersion, (3) applying shear or other forces to form droplets of the internalphase and reduce the mean droplet size, and (4) allowing sufficient time inthe shear zone to adequately cover the internal phase with emulsifier tostabilize the droplets. With the correct emulsifier, food emulsions can beproduced with very little mechanical energy input; dressings and sauces havebeen prepared for hundreds of years using no more than a bowl and a whisk.However, the use of more complex emulsification devices enables productattribute optimization at the lowest cost. Generally, dressings and saucesthat are dispersions as opposed to emulsions (typically low- and no-fatdressings and sauces) may be produced on the same process that is usedto produce emulsions, although different considerations may take priority,such as the shear rate needed to create a uniform product in the case of adispersion being of greater importance than the amount of time in the shearzone necessary for emulsifier adsorption in the case of an emulsion.

A. Drop Breakage

Drop breakage occurs in emulsification devices under laminar, turbulent,or cavitation modes, or combinations thereof. Elongational flow isencountered under some conditions. For an excellent account of the dropbreakage under these various conditions, the reader is referred to Chapter 1of Ref. 12.

Studies carried out by Taylor (72) on single droplets under laminarconditions have shown that the drop undergoes steady deformation undershear up to a point at which breakage can occur into smaller fragments. Theextent of deformation is characterized by a dimensionless number called theWeber number, WeL, which is defined as the ratio between the deformingshear forces and conservative interfacial forces, which tend to restore thespherical, undeformed shape of the drop:

WeL ¼�

pc¼�d

4�ð2Þ

Here, � is the shear stress, pc is the capillary pressure, g is the inter-facial tension and d is the drop diameter. At a certain critical Webernumber, a critical deformation state is reached at which rupture occurs.


The critical Weber number is a function of the viscosity ration of the twophases:

WeLcrit ¼ f�d�c

� �

ð3Þ

Figure 8 shows the variation of Wecrit with the viscosity ratio. Thus, themaximum size of the drop that can withstand rupture is given

dmax ¼4�WeLcrit

�ð4Þ

Equation (4) shows that selection of both the proper emulsifier to lowerinterfacial tension and process to provide a high shear rate will aid inachieving a small dmax. It is also clear then that in order to cause dropletbreakage under laminar conditions, the maximum droplet size that cansurvive rupture depends strongly on the viscosity ratio. A viscosity ratioof near unity gives the smallest droplet sizes for a given shear input and isthe most energy-efficient situation.

Turbulent flow is characterized by eddies of a wide size range. Overallliquid movement is due to the large eddies of the order of the energy-trans-mitting device such as the agitator. These eddies transfer kinetic energy tothe small eddies where viscous dissipation occurs. If a droplet in turbulent

Figure 8 Variation of critical Weber number with viscosity ratio. (From Ref. 75.)


flow is much smaller than the eddy, it follows the movement of the eddy. If,however, it is of a size similar or larger than the eddy, fluctuating velocitygradients at the droplet surface cause droplet deformation, which can leadto rupture. In analogy to the laminar conditions, a turbulent Weber numbercan be defined as the ratio between the deforming turbulent forces and therestoring interfacial forces:

WeT ¼�cv

02d

4�ð5Þ

where �c is the continuous phase density and v0 is the eddy velocity.According to the Kolmogoroff theory of isotropic turbulence, we get

WeT /�c"

2=3d5=3

�ð6Þ

Rupture, again, occurs when the Weber number exceeds a critical value:

WeT >WeTcrit ð7Þ

Thus,

dmax /�3=5

�3=5c "2=5ð8Þ

Equation (8) is valid for dmax larger than the size of the small energy-dis-sipating eddies, which is commonly the case in a simple emulsification devicewithout homogenization facilities.

When the viscosity of the system is large or the drops are very small,Shinnar (73) has suggested that the drop breakage even under highly turbu-lent conditions occurs due to viscous stresses rather than the inertial stresses,and

dmax /�2

�c�c"

� �1=2

ð9Þ

This equation is not fully tested experimentally, perhaps because the high-pressure homogenization condition that are perhaps necessary to achievesmall drops involve multiple mechanisms and not simple turbulent shear fordrop breakage.


Finally, cavitation is the phenomenon of formation and collapse ofsmall vapor bubbles in a liquid. A high-velocity fluid produces locally nega-tive pressures which lead to the formation of a cavity at that point. As thecavity implodes, it produces a macroscopic shock were. A nearby dropletcan get sucked into the shrinking void and the resultant droplet deformationmay lead to its rupture. Cavitation is implicated in ultrasonic emulsificationand, perhaps, in a high-pressure homogenization, although Phipps (74), hasrejected this notion.

The above equations describing drop breakage assume that the dropstays in a high-energy zone for sufficiently long time for it to deform andbreak. This critical deformation time depends on the drop viscosity and theexcess deforming stress above the conservative interfacial stress

tdef ¼�d

� � pcð10Þ

For a drop size reduction by a factor of 10, approximately 10 consecutivedisruption steps are necessary, assuming that each disruption step results intwo drops of size (d/2)1/3. The overall residence time in the high-energy zonemust be longer than the sum of individual critical deformation times.The influence of the residence time is particularly evident under turbulentconditions.

B. Emulsifier Effect

Given that the emulsifier is capable of stabilizing the emulsion in the firstplace, the conditions needed to be met to minimize recoalescence are that theemulsifier is present in sufficient quantity to adsorb on the newly formedsurfaces to provide minimally a monolayer coverage and that the kinetics ofadsorption are rapid. The approximate monolayer coverae for low-molecu-lar-weight lipid based emulsifiers such as polysorbates is of the order of5� 10�10mol/m2 and for macromolecular emulsifiers such as proteins it isof the order of 2–8mg/m2. Based on the interfacial area generated, one cancalculate the amount needed for adsorption. The adsorbed emulsifier is inequilibrium with the emulsifier in the bulk. If the adsorption isotherm isknown, one can calculate precisely the minimum amount of emulsifierneeded. Otherwise, one can add a little excess over the amount needed atthe interface.

Adsorption kinetics are governed primarily by the size of the emulsi-fier molecule. When the emulsifier adsorbs rapidly, the situation is what iscalled a mechanically limited dispersion. Here, the above drop breakageequations apply, with g being the equilibrium interfacial tension at theemulsifier concentration involved (this approaches gmin, the minimum


achievable interfacial tension for the emulsifier in question if the emulsifieris present in sufficient excess). When the emulsifier adsorbs slowly, theinterfacial tension in the above equations is not the equilibrium interfacialtension at the emulsifier concentration involved but is, in fact, quite a bithigher, resulting in a larger drop size. This situation is described as kineti-cally limited dispersion. Furthermore, without sufficient adsorption of theemulsifier, the recoalescence process is not completely prevented, furtherincreasing the droplet size obtained. Under laminar conditions, the largeremulsifier molecules show slow adsorption kinetics. Figures 9 and 10 showthe droplet sizes obtained for a fast emulsifier LEO-10, lauryl alcohol10(ethylene oxide) ether, and a slow emulsifier, egg yolk (75).* Thesedata are the first such data known to the authors showing the effect ofadsorption kinetics of emulsifier on the final droplet size obtained duringemulsification.

Although it is understandable that the larger emulsifier moleculesshow slow adsorption kinetics under laminar conditions, Walstra (65) hassuggested that the exact opposite may be applicable under turbulent condi-tions, especially when the emulsifier molecular sizes approach the size of thedrops. The phenomenon of capture during collisions is then likely involved.Capture efficiencies are expected to be significantly higher for such largeemulsifier molecules. No experimental data similar to those of Armbrusterare available under turbulent conditions to verify this assertion.

*The lines in the figures are theoretical calculations based on Eq. (4) with the area average

diameter estimated to be dmax/2.4 for critical Weber numbers estimated from the viscosity

ratios (not shown).

Figure 9 Droplet diameter as a function of shear stress for a fast emulsifier,

LEO-10. (From Ref. 75.)


The Weber relationships discussed earlier suggest that droplet size maybe reduced by (a) reducing interfacial tension and by (b) increasing processshear rate in the case of laminar flow or fluid velocity in the case of turbu-lent flow. However, following the above discussion, depending on whether asystem is a kinetically or mechanically limited there may be limits on dropletsize reduction available in a given system that will drive process developer tomodify the selection and operating procedures of emulsification equipment.For example, it is possible to show that when using a homogenizer toprocess a slow emulsifier system at a constant emulsifier level, there existsan inverse relationship between homogenization pressure and droplet sizefor low-fat systems such that processing at progressively higher pressureswill deliver progressively smaller droplets. However, using the same processand emulsifier systems at a higher fat level, progressively higher pressureswill initially yield smaller droplets, but an asymptote will be reached wherehigher pressures will not reduce droplet size further. Although the initialdroplet size may be the same at high pressures for the different fat levels, theextremely short residence time in the homogenizer limits the extent of emul-sifier adsorption in the shear zone. Therefore, recoalescence is more likely tooccur in the high fat/high surface area system resulting in an asymptotic

Figure 10 Droplet diameter as a function of shear stress for a slow emulsifier, egg

yolk. (From Ref. 75.)


relationship between particle size and homogenization pressure such that nofurther droplet size reduction is possible beyond a certain pressure. If smal-ler droplets were desired in a high-fat system, it would be necessary to eitherchange the type of emulsifier or equipment to make the system mechanicallylimited.

C. Process Design and Equipment Selection

Important considerations in designing a food emulsion process include pro-cess type (batch versus continuous), order of addition of ingredients, processtemperature, and process shear rates. The decision of whether to use a batchor continuous process is driven by consideration of run-time length, space,frequency of manual ingredient additions, and capital requirements.Processes which have a high number of product changeovers and/or a sig-nificant number of ingredient additions are ordinarily designed as batchprocesses. Continuous processes are typically used in cases of long runtimes and offer advantages of reduced labor and space requirements.

The selection of the final emulsification device that will achieve theminimum droplet size and produce the finished product is of great impor-tance. Although the feed to this device may consist of either a coarse emul-sion or an unstable dispersion, a coarse emulsion is often used. The benefitof the coarse emulsion is the partial reduction in droplet size and enhanceduniformity of feed to the final emulsification device. When preparing thefeed to this unit operation, a coarse emulsion may be obtained using anemulsifier soluble in the continuous phase and adding the discrete phase tothe continuous phase (typically, the discrete phase will be oil and the con-tinuous phase will be aqueous). In a batch process, a mixing tank may beused prior to the shear device to provide the coarse emulsion. A variety ofagitator types may be found in the industry, including axial-, radial-, anddisk-type impellers. In continuous or semicontinuous processes, the oil andaqueous streams are joined and sheared via in-line mixers to form the coarseemulsion prior to the final emulsification device. Many in-line mixes are alsocommercially available for this application.

Although many choices are available for the emulsification unit opera-tion, they generally fall into two categories: (1) devices that reduce dropletsize via shear forces in laminar flow and (2) devices that reduce droplet sizevia cavitation and/or shear in turbulent flow (a number of devices employboth methods). Shear forces result from velocity gradients; most sheardevices generate shear forces by passing fluid at high velocities throughsmall stationary or moving gaps or by passing fluid from a region of highto low pressure. In general, a higher shear rate will produce a smaller andmore uniform droplet size, although there is a point of diminishing returns


and the temperature rise at very high shear rates may be detrimental to theproduct. Commonly used shear devices include colloid mills, toothed-diskrotor/stator mixers, and pin mixers. Colloid mills utilize a single conicalrotor–stator pair in which the rotor surface geometry and the gap betweenthe rotor and stator may be adjusted to vary shear rate. Tooted-disk rotor–stator devices generally consist of three or more sets of rotors/stators inwhich the tooth geometry, rotor/stator gap size, and rotor may be adjustedto modify shear rate. As discussed earlier, cavitation refers to the implosionof vapor bubbles within a fluid; the resultant shock waves create droplets.Shear generated in turbulent flow may contribute to droplet formation andsize reduction due to intense fluid mixing and localized pressure differences.The most common example of this method is homogenization in whichvapor bubbles are formed by pressure differential in fluid flow.

The choice of equipment may be driven by formulation. As discussedearlier, products using slow emulsifiers and high oil levels may be bettersuited to equipment which provides sufficient residence time to allow fordroplet stabilization, whereas products using fast emulsifiers and/or low-fatlevels may be better suited to equipment such as a homogenizer, whichprovides significant droplet size reduction with very little residence time inthe unit (although for all equipment, residence time is on the order ofseconds).

For oil-based dressings and sauces, attention must also be paid tominimizing oxidation of the oil content through the processing and packa-ging operations. In addition to formulation methods including the use ofsequestrants and antioxidants, common methods include processing in aninert-gas environment and the use of barrier packaging materials to displaceoxygen from the product and prevent its reintroduction.

Table 6 lists some of the product-impact parameters that should beconsidered when choosing an emulsification device. Manufacturing consid-erations such as flexibility in modifying shear rates, sanitary design, CIPability, energy usage, maintenance costs, changeover flexibility, and ergo-nomics in disassembly must also be considered and may often define thechoice between equipment that has similar effects on product attributes.

It is a common misconception to assume that the central role of anemulsification device is to provide energy input needed because one isincreasing the interfacial area, and hence, the interfacial energy of thesystem. In reality, only a small fraction of the energy input actually goestoward the interfacial energy. Most of it is dissipated as heat. Estimates arethat only 0.01% of the energy used in a homogenizer is converted to theinterfacial energy. That is why many of the high-energy emulsificationdevices cause a significant temperature rise in the emulsion, sometimesnecessitating a cooling scheme. In fact, a good estimate of the energy


Table 6 Criteria of Emulsification Equipment Selection

Type of machine Agitated vessel Colloid mill Toothed disk dispersers Homogenizers

Communition Turbulent shear Laminar turbulent shear Turbulent shear Turbulent shear/cavitation

mechanisms

Energy 103–105 106–1011 108–109 1011–1013

consumption (W/m3)

Residence time (s) Undefined 10�3–10�1 10�3–10�2 10�5–10�4

Droplet size (mm) 5–100 2–20 2 –20 0.5–5

Suitable for high-fat

o/w emulsions

þ þ �

Suitable for low viscosities � þ

Suitable for high viscosities þ þ �

Addition of particulates þ � �

Indirect heating/cooling þ �

Direct steam injection þ �

Continuous processing � þ þ þ

Batch processing þ þ þ �


input by an emulsification device can be made from the temperature riseobtained if no cooling is provided or from a heat balance if cooling isprovided. In the case of a homogenizer, for every 1000 psi homogenizingpressure, the temperature rise is approximately 3� F.

VI. NONCLASSIC DRESSINGS AND SAUCES

In previous sections, the full-fat versions of traditional emulsion-based pro-ducts have been presented and, in this chapter, will be referred to as classicemulsions. In this section, nonclassic emulsions will be considered as no-fatto low-fat versions of standard full-fat products.

On a chronological basis, spoonable salad dressings developed in the1930s can be considered as a first-generation, reduced fat mayonnaiselikedressing. Oil levels were reduced from 80% in mayonnaise to 45–50% insalad dressing. With this substantial reduction in dispersed phase volume,significantly thinner texture was supplemented through the use of starches.Flavor changes were masked through targeting a tarter, sweeter, and morespicy profile.

During the mid-1980s, ‘‘light’’ dressings were introduced. Gums wereincluded in formulations to add texture such as mouth-feel, cling, firmness,slipperiness, and so forth that could not be provided in an acceptablemanner through merely increased levels of starch.

Then, during the early 1990s, fat-free dressings were developed. Inmost cases, starches and gums used alone or in conjunction with eachother were not able to deliver the texture and flavor necessary to gainsatisfactory consumer acceptance. Fat replacers or fat mimetics were thetechnologies developed for the most part to provide the additional textureand appearance characterisits needed.

The progression of fat/emulsion phase reduction is readily apparent ina microscopic investigation. The loss of a significant portion of the fat in themove to fat free is evidenced in the example of spoonable salad dressings inwhich the increasing role of starch and fibrillarlike fat replacers are shownas the fat/emulsion phase is reduced from full fat (48%), light (23%), and fatfree (2–3%) (Figs. 11–11c).

Earliest fat replacers were built on the concept of non-digestible fatlike molecules. Olestra from Procter and Gamble was developed in the 1960sfrom sugar and fatty acids to produce sucrose hexa, hepta and octa fattyacid esters. This fat replacer was designed to replace oil on a one-to-onebasis, functioning as the hydrophobic phase in emulsions and even capableof fried and cooked applications. However, its non digestibility and macroingredient status raised issues concerning fat-soluble vitamin depletion and


the body’s ability to handle it as waste. A number of similar products havebeen developed such as esterified polysaccharides, carboxylate esters, and soforth by various companies but none of them are in use in dressings andsauces.

During the late 1980s a paradigm shift occurred such that rather thantrying to replace all of the functionality of fat with a fatlike molecule, ofperhaps some of the functionality of fat could be substituted through micro-particulates. The original product based on this concept was Simplesse fromthe NutraSweet Company. Using egg and dairy protein as source materials,a creamy material was produced which was composed of spherical protein

Figure 11 The microstructure of spoonable salad dressing. (a) Section of stained,

full-fat sample viewed with bright-field light microscopy. Lipid droplets (l) are

numerous and are separated by some starch granules. Scale bar equals 25 mm. (b)

Section of stained, ‘‘light’’ sample viewed with bright-field light microscopy. Lipid

droplets (l) are significantly less numerous and are separated by more starch

granules. Scale bar equals 25 mm. (c) Section of stained, fat-free sample viewed with

bright-field light microscopy. Lipid droplets (l) are scant and are separated by starch

granules and fibrillarlike material. Scale bar equals 25 mm.


particles in the 0.5–2.0-mm diameter range. It was claimed that the particu-lates produced a ball-bearing effect in the mouth. Later, larger particles werealso found to work if soft and compressible. This concept was soon trans-ferred to carbohydrate-based materials and their gels. Starch and gum tech-nologies were reevaluated and novel textures were developed to supplementconventional usages. It was this paradigm shift in the 1980s that led to thesuccessful introduction of fat-free products.

The original approach for replacing the functionality of the fat fromthe emulsion phase was to search for a ‘‘magic bullet’’ ingredient whichwould target all the product needs. This approach was only somewhatsuccessful in products which fat contributed in fairly straightforward andminor ways to texture and where the product had strongly characterizingflavors which were mostly independent on fat contribution or interaction. Asystems approach was soon adopted to replace specific product functional-ity of the fat by specific fat-replacement technology. The impact of fat inthe emulsion phase is evidenced in the product in appearance, texture,mouth-feel, stability, handing, and, most importantly, flavor (Fig. 12).

Flavor character, release, and stability continued to be a major hurdlein the nonclassic emulsion dressings. Fat is known to play an important rolein the flavor perception of foods. It influences the temporal profile, flavorimpact, perception of flavor notes, and the order of their occurrence (76).The fat replacers in use today are composed of proteins and carbohydrates,which interact with the flavor compounds differently than fat does (77,78).For example, flavors are known to bind to proteins (79) and starches (80).

Figure 12 Functional considerations in the development of fat free products.

(Adapted from Prepared Foods, December 1992, p. 21.)


The influence of the oil phase in the emulsion of classic emulsions on theproduction, perception, and preservation of flavors was considerable andpresented many challenges for replication in nonclassic dressings. The nearabsence of fat in fat-free products posed a difficult challenge in achievingfull-fat flavor (Table 1). An understanding of these flavor effects nowbecame necessary.

The overwhelming majority of food flavors are complex mixtures ofcompounds. These flavors come to dressings through ingredients, productprocessing during manufacturing, flavor addition, chemical reactions duringstorage, physical changes including repartitioning of and/or loss of phases,and final food preparation.

Preception of flavor depends not only on the presence of proper flavorcompounds at the appropriate intensity but the timing of the flavor releaseand masking effects as well. In nonclassic emulsions, the increased presenceof water and the hydrophilic molecules of starch, gum, and other fat repla-cers dramatically affects the partitioning of volatile hydrocarbon flavorcompounds. This effect significantly affects flavor intensity perception andrelease rates. The masking effect of the dispersed oil in the emulsion ofclassic emulsions plays an important role in rounding flavor perceptionand covering low-level off-flavors. Also, the presence of fat affects the bind-ing of flavors by various hydrophobic sites in carbohydrates and proteins.Quality gaps in fat-free products along with low regulatory hurdles haverecently opened the door for reduced-calorie fats. Products such asCaprenin and Salatrim are special triglycerides based on specific fatty acidprofiles which digest at lower caloric values. These products could allow foran intermediate approach of using lower levels (but not fat free) of a fatingredient at a significantly reduced calorie impact.

VII. CONCLUSIONS AND PROGNOSTICATION

There are four significant trends in the dressings and sauces industry. Firstis the introduction of increasing number of low-fat and fat-free dressingsand the continuing effort to improve these products. Second is a steadylaunch of new products, especially new flavors, especially in sauces andpourable dressings categories. Third is the growing need to cut the cost ofproduction. Fourth are next trends of healthy foods—for example, basedon new fats and oils such as diacyl glycerols (DAGs) such as Enova oilrecently introduced in the market by ADM Kao. Accomplishing this willinvolve changes in formulation and processing. Developing products scien-tifically would necessitate application of disciplines such as rheology,microscopy, and emulsion stability theory; the objective data provided


by such disciplines will have to be correlated with sensory evaluation ofthe products.

Particularly critical is the advancement of fundamental understandingof emulsion stability. Successful predictive tests for emulsion stability cansignificantly reduce the product development cycle time. We believe herethat the role of thin films in the stability of emulsions is an emerging field ofstudy that is likely to be the key to predicting emulsion stability of dressingsand sauces over the shelf life as well to their processing.

Microscopy of foods has come a long way in recent years. For arecent review of available techniques, see Ref. 81. Significant technicaladvances in the field of microscopy include cryo-fixation methods, whichwill replace chemical fixation with a reduction in processing time andelimination of artifacts. Expanded applications will be made of immuno-cytochemical methods to determine the location of target proteins invarious food systems. Computer-aided image analysis and enhancementsystems will allow more objective, rapid, and accurate analysis of micro-scopic data than current morphometric and stereological methods.Application will be made of confocal microscopy and the ‘‘optical section-ing’’ capability of that technique. The scanning tunneling microscope [andthe atomic force microscope (AFM)] will be used to resolve structure at amolecular level.

Advances in theory and instrumentation have allowed rigorous viscoe-lastic characterization of food products. As discussed, newer techniques arebeing developed to characterize extensional properties and tribology of theproducts.

With better microscopic and rheological techniques, it will be possibleto relate microstructure to rheology and, finally, sensory textural character-istics of the product, a goal which has heretofore remained elusive. Thesetechniques, and others, will be used to understand and predict the emulsionstability of dressings and sauces.

REFERENCES

1. N. N. Potter, Food Science, 4th ed. AVI, Westport, CT, 1986, pp. 457–458.

2. C. M. Chang, W. D. Powrie, and O. Fennema, Electron microscopy of

mayonnaise, Can. Inst. Food Sci. Technol. J. 5, 134–137 (1972).

3. M. Tanaka and H. Fukuda, Studies on the texture of salad dressings containing

xanthan gum, Con. Inst. Food Sci. Technol. J. 130–134 (1976).

4. M. A. Tung and L. J. Jones, Microstructure of mayonnaise and salad dressing,

Scanning Electron Micros. 3 523–530 (1981).

5. W. Flukiger, Die fliesseigenschaften von emulsonen, Fette-Seifen Anstrichmittel

69, 449–452 (1967).


6. D. F. Darling and R. J. Birkett, Food colloids in practice, in Food Emulsions

and Foams (E. Dickinson, ed.), Royal Society of Chemists, London, 1986, 1–29.

7. I. Heertje, P. van der Vlist, J. C. G. Blonk, H. A. C. Hendricks, and

G. J. Brakenhoff, Confocal scanning laser microscopy in food research: Some

observations, Food Microstruct. 6, 115–120 (1987).

8. R. K. Salyaev, A method of fixation and embedding of liquid and fragile

materials in agar microcapsulae, in Proc. 4th European Regional Conf. Electron

Microscopy, 1968, pp. 37–38.

9. M. Kalab, Electron microscopy of milk products: A review of techniques.

Scanning Electron Microsc. 3, 453–472 (1981).

10. P. Allan-Wojtas and M. Kalab, Fixation of fat globules in whole milk yogurt

for electron microscopy, Milchwissenschaft 39, 323–327 (1984).

11. M. Liboff, R. D. Goff, Z. Haque, and J. E. Kinsella, Changes in the

ultrastructure of emulsions as a result of electron microscopy preparation

procedures, Food Microstruct. 7, 67–74 (1988).

12. P. Sherman (ed.), Emulsion Science, Academic Press, New York, 1968.

13. S. D. Holdsworth, Applicability of rheological models to the interpretation of

flow and processing behavior of fluid food products, J. Texture Studies, 2,

393–418 (1971).

14. O. Flint, Food emulsions, in Food Microscopy A Manual of Practial Methods,

Using Optical Microscopy, Bios Scientific, 1987, pp. 113–118.

15. L. Munck (ed.) Fluorescence Analysis in Foods, Wily New York, 1989.

16. T. Sengupta, and S. Damodaran, Lateral phase separation in adsorbed binary

protein films at the air-water interface, J. Food Chem. 49 (6), 3087–3091 (2001).

17. I. Heertje, J. Nederlof, H. A. C. M. Hendricks, and E. H. Lucassen-Reynders,

The observation of the displacement of emulsifiers by confocal scanning laser

microscopy, Food Struc. 9, 305–316 (1990).

18. M. Langton, A. Astrom, and A.-M. Hermansson, Microscopy in relation to the

textural properties of mayonnaise, in Food Emulsions and Foams: Interfaces,

Interactions and Stability RSC, Cambridge, 1999, pp. 366–376.

19. R. P. Borwankar, Food texture and rheology: A tutorial review, J. Food Eng.

16, 1–16 (1992).

20. F. Shama and P. Sherman, Identification of stimuli controlling the sensory

evaluation of viscosity. I. Oral methods. J. Texture Studies, 111–118 (1973).

21. J. L. Kokini, Fluid and semi-solid texture and texture taste interactions, Food

Technol. 39(11), 86–94 (1985).

22. R. L. McDermott, W. J. Harper, and R. Whitley, A centrifugal method for

characterization of salad dressing emulsions, Food Technol. 98, 81–87 (1981).

23. G. Atkin and P. Sherman, Further application of the modified gel rigidity

modulus apparatus, J. Texture Studies 10, 253–259 (1980).

24. R. Fischbach and J. L. Kokini, Effect of aging and mustard flour on

rheological properties of model O/W emulsion, J. Food Sci. 52, 1748–1749

(1987).

25. V. D. Kiosseoglou and P. Sherman, The influence of egg yolk lipoproteins on

the rheology and stability of o/w emulsions and mayonnaise. 1. Visco-elasticity


of groundnut oil-in-water emulsions and mayonnaise, J. Texture Studies 14,

397–417 (1983).

26. J. I. Lewis and C. F. Shoemaker, A method for measurement of the transient

response of semisolid foods under steady shear, J. Food Sci. 49, 741–755 (1984).

27. E. J. Vernon-Carter and P. Sherman, Rheological properties and applica-

tions of mesquite tree (Prosopis juliflora) gum. 4. Rheological properties of

mesquite gum films at the oil water interface, J. Dispers Sci. Technol. 2, 399–413

(1981).

28. E. J. Vernon-Carter and P. Sherman, Rheological properties and applications

of mesquite tree (Prosopis juliflora) gum. 5. The rheological properties and

stability of corn oil-in-water emulsions stabilized by blends of mesquite gum

and Tween 60, J. Dispers Sci. Technol. 2, 415–431 (1981).

29. K. K. Bistany and J. L. Kokini, Comparison of steady shear rheological

properties and small amplitude dynamic visco-elastic properties of fluid food

materials, J. Texture Studies 14, 113–124 (1983).

30. M. Dervisoglu and J. L. Kokini, Steady shear rheology and fluid mechanics of

four semi-solid foods, J. Food Sci. 51, 541–546 (1986).

31. J. H. Elliott and A. J. Ganz, Salad dressings—Preliminary rheological

characterization, J. Texture Studies 8, 369–371 (1977).

32. J. L. Kokini and A. Dickie, A model of food spreadability from fluid

mechanics, J. Texture Studies, 13, 211–227 (1981).

33. J. A. Blake and J. J. Morn, A new approach to capillary extrusion rheometry,

J. Texture Studies 5, 227–239 (1975).

34. O. H. Campanella and M. Peleg, Analysis of the transient flow of mayonnaise

in a coaxial viscometer, J. Rheol. 31, 439–452 (1987).

35. O. H. Campanella and M. Peleg, Determination of the yield stress of semiliquid

foods from squeezing flow data, J. Food Sci. 52, 214–215 (1987).

36. V. D. Kiosseoglou and P. Sherman, The rheological conditions associated with

judgment of pourability and spreadability of salad dressings, J. Texture Studies

14, 277–282 (1983).

37. P. Kovacs and B. D. Titlow, Stabilizing cottage cheese creaming emulsions with

a xanthan blend, Am. Dairy Rev. 34J–34M (1976).

38. N. Gladwell, M. J. Grimson, R. R. Rahalkar, and P. Richmond, Rheological

behavior of soy oil-water emulsions: Dependence upon one concentration,

J. Food Sci. 60, 440–443 (1985).

39. N. Gladwell, M. J. Grimson, R. R. Rahalkar, and P. Richmond, Linear visco-

elastic behavior of sterically stabilized oil-water emulsions, J. Chem. Soc.,

Faraday Trans 2 81, 643–652 (1985).

40. M. A. Rao, Measurement of flow properties of fluid foods—developments,

limitations, and interpretation of phenomena, J. Texture Studies, 8, 257–282

(1977).

41. Boger and Tiu 1974.

42. C. Pascual, M. C. Alfaro, and J. Munoz, Food Emulsions and Foams: Interfaces,

Interactions and Stability. E. Dickinson and J. M. Rodriguez Patino, eds.),

RSC, Cambridge, 1998, pp. 356–365.


43. H. A. Barnes and K. Walters, The yield value myth? Rheol. Acta 24, 323–326

(1985).

44. G. W. Scott Blair, J. Appl. Phys. 4, 113 (1933).

45. N. J. Alderman, G. H. Meeten, and J. D. Sherwood, Vane rheometry of

bentonite gels, J. Non-Newtonian Fluid Mech. 39, 291–310 (1991).

46. N. A. Park, R. G. Irvine, Jr., and F. Gui, Yield stress measurements with the

falling needle viscometer, Xth Int. Cong. on Rheology, 1988.

47. N. A. Park and T. F. Irvine, Jr., The falling needle viscometer: A new technique

for viscosity measurements. Am. Lab. 58–63 (1988).

48. J. Plucinski, R. K. Gupta, and S. Chakrabarti, Shear and extensional rheology

of mayonnaises Second International Meeting of Shear and Extensional Flow of

Polymer Fluids from the Solution to the Melt, 1994.

49. S. Giasson, J. Israelachvili, and H. Yoshizawa, Thin film morphology and

tribology of food emulsions J. Food Sci. 62, 640–646 (1997).

50. P. Sherman (ed.), Rheology of Emulsions, Pergamon Press, Oxford, 1963.

51. P. Sherman (ed.), Food Texture and Rheology, Academic Press, New York,

1978.

52. R. J. Hunter, Introduction to Modem Colloid Science, Oxford Science, Oxford,

1993.

53. C. F. Chan Man Fong, G. Turcotte, and D. De Kee, Modelling steady and

transient rheological properties, J. Food Eng. 27, 63–70 (1996).

54. J. I. Amemiya and C. F. Shoemaker, Measurement of thixotropy of model food

colloidal suspensions with step change shear rate, in Rheology of Foods (R. P.

Borwankar and C. F. Shoemaker, eds.), Elsevier, New York, 1992, pp. 17–24.

55. H. M. Princen, Rheology of foams and highly concentrated emulsions. J.

Elastic properties and yield stress of a cylindrical model system. J. Colloid

Interf. Sci. 91, 160–175 (1983).

56. I. B. Ivanov, Effect of surface mobility on the dynamic behavior of thin liquid

films. Pure Appl. Chem. 52, 1241–1261 (1980).

57. I. B. Ivanov and D. S. Dimitrov, Thin film drainage, in Thin Liquid Films

(I. B. Ivanov, ed.), Marcel Dekker, New York, 1988.

58. D. T. Wasan, Interfacial transport processes and rheology: Structure and

dynamics of thin liquid films. Chem. Eng. Ed. 104, (1992).

59. A. D. Nikolov and D. T. Wasan, Ordered micelle structuring in thin films

formed from anionc surfactant solutions: part I—Experimental. J. Colloid

Interf. Sci. 133, 1–12 (1989).

60. A. D. Nikolov, P. A. Krachevsky, I. B. Ivanov, and D. T. Wasan, Ordered

micelle structuring in thin films formed from anionic surfactant solutions:

Part II—Model development. J. Colloid Interf. Sci. 133, 13–22 (1989).

61. K. Koczo, A. D. Nikolov, D. T. Wasan, R. P. Borwankar, and A. Gonsalves,

Layering of sodium caseinate sub-micelles in thin liquid films. J. Colloid Interf.

Sci. 178(2), 694–702 (1996).

62. S. U. Pickering, J. Chem. Soc. 91, 2001 (1907).

63. W. D. Harkins and N. Sollman, Interfacial tension and emulsification.

I. The effects of bases, salts, and acids upon the interfacial tension


between aqueous sodium oleate solutions and benzene, J. Am. Chem. Soc. 48,

69 (1926).

64. P. Graham, The conformation of proteins at interfaces and their role in

stabilizing emulsions, in Theory and Practice of Emulsion Technology (A. L.

Smith, ed.), Academic Press, New York, 1976, pp. 75–98.

65. P. Walstra, Overview of emulsion and foam stability, in Food Emulsions and

Foams (E. Dickinson, ed.), Royal Society of Chemists, London, 1987.

66. W. D. Powrie and S. Nakai, The chemistry of eggs and egg products in Egg

Science and Technology (W. J. Stadelman and O. J. Cotterill, eds.), Food

Products Press, 1990, pp. 97–139.

67. R. E. Baldwin, Functional properties of eggs in foods in Egg Science and

Technology (W. J. Stadelman and O. J. Cotterill eds.), Food Products Press,

1990, pp. 345–383.

68. T. M. Riddick, Control of Colloid Stability Through Zeta Potential. Zeta-Meter,

Inc., 1968, p. 150.

69. J. W. Corran, Emulsification by mustard, Spice Mill 57, 175 (1934).

70. B. H. Webb, A. H. Johnson, and J. A. Alford, Fundamentals of Dairy

Chemistry, 2nd ed. AVI Westport, CT, 1983, p. 90.

71. P. Walstra, Formation of emulsions, in Encyclopedia of Emulsion Technology,

Vol. 1: Basic Theory, (P. Becher, ed.), Marcel Dekker, New York, 1983.

72. G. I. Taylor, Proc. R. Soc. London A 146, 501 (1934).

73. R. Shinnar, On the behavior of liquid dispersions in mixingvessels, J. Fluid

Mech. 10, 259–275 (1961).

74. L. Phipps, The High Pressure Dairy Homogenizer, National Institute for

Research in Dairying, Reading, PA, 1985.

75. H. Armbruster, Untersuchungen zum kontinuierlichen Emulgierprozess in

Kolloidmuhlen unter Berucksichtigung spezifischer Emulgatoreigenscheften

und Stromungsverhaltnissen im Disperfierspalt. Dissertation, University of

Karisruhe, Karlsruhe, Germany.

76. L. C. Hatchwell, Implication of Fat on Flavor in Flavor–Food Interactions (R. J.

McGorrin and J. V. Leland, eds.), ACS Symposium Series 633, American

Chemical Society, Washington, DC, 1996, pp. 14–23.

77. J. Bakker, Flavor interactions with the food matrix and their effects on

perception in Ingredient Interactions (A. G. Gaonkar, ed.), Marcel Dekker,

New York, 1995, 411–439.

78. G. A. Reineccius, Flavor fat replacer Interactions in foods, in Ingredient

Interactions (A. G. Gaonkar, ed.), Marcel Dekker, New York, 1995,

pp. 441–450.

79. T. E. O’Neill, Flavor Binding by Food Proteins: An Overview in Flavor–Food

Interactions (R. J. McGorrin and J. V. Leland, eds.), ACS Symposium Series,

633 American Chemical Society, Washington, DC, 1996, pp. 59–74.

80. M. Y. M. Hau, D. A. Gray, and A. J. Taylor, Binding of volatiles to Starch in

Flavor–Food Interactions (R. J. McGorrin and J. V. Leland, eds.), ACS

Symposium Series 633 American Chemical Society, Washington, DC, 1996, pp.

109–117.


81. M. G. Smart, R. G. Fulcher, and D. G. Pechak, Recent developments in

the microstructural characterization of foods, in Characterization of

Foods: Emerging Methods (A. G. Gaonkar, ed.), Elsevier, New York, 1995,

pp. 233–275.

82. J. Sestak, R. Zitny and M. Houska, Simple rheological models food liquids for

process design and quality assessment, J. Fodo Eng. 35–49 (1983).


14Analysis of Droplet CharacteristicsUsing Low-Intensity Ultrasound

John N. Coupland

The Pennsylvania State University, University Park,Pennsylvania, U.S.A.

D. Julian McClements

The University of Massachusetts, Amherst, Massachusetts, U.S.A.

I. INTRODUCTION

A. Introduction

Sensory and bulk physicochemical properties of food emulsions, such astexture, stability, appearance, and taste are largely determined by thecharacteristics of the droplets that they contain (i.e., particle size distribu-tion, disperse phase volume fraction, physical state, and colloidal interac-tions) (1). For example, the rate at which droplets cream or sediment dueto gravitational forces is strongly dependent on their size. The character-istics of the droplets in a particular emulsion are governed principally bythe ingredients it contains, the homogenization technique used to prepareit, and the environmental conditions it experiences during manufacture,storage, and consumption and may change appreciably during the lifetimeof the product due to various instability mechanisms. For these reasonsand because it is impossible to accurately predict the structure/function-ality for many real systems, it is important to have analytical techniquesthat can be used to measure the characteristics of the droplets inemulsions.


The compositional and microstructural complexity of food emulsionsoften makes traditional methods of droplet characterization unreliable(e.g., microscopy, light scattering, and electrical pulse counting) Con-sequently, the development of novel methods of characterizing dropletcharacteristics in food emulsions has gained impetus in recent years. Theideal analytical technique for measuring droplet characteristics would benon-destructive, simple to use, versatile, rapid, reproducible, and reasonablypriced. It would also be valuable to be able to make measurements onemulsions during the manufacturing process, so the technique shouldmeet hygienic processing standards and be robust enough to survive in afactory environment. Analytical techniques based on ultrasound meet manyof these requirements.

B. Ultrasonic Propagation in Food Emulsions

Analytical techniques based on ultrasound use high-frequency mechanicalvibrations (typically between 20 kHz and 200MHz) to provide informationabout the composition, structure, or dynamics of materials (2). Ultrasonicwaves are qualitatively similar to sound waves, but their frequencies are toohigh for the human ear to detect. Ultrasonic waves propagate throughmaterials as small deformations in the thermal–mechanical properties (i.e.,as small deviations in the average temperature and pressure of the material).The power levels used in the ultrasonic analysis of food emulsions are so lowthat the deformations caused in the material are extremely small and rever-sible, which means that the technique is nondestructive.

In general, the most commonly used ultrasonic waves for characteriz-ing materials are longitudinal waves and shear waves. In longitudinal waves,the deformations occur parallel to the direction of transmission of the wave,whereas in shear waves, the deformations occur perpendicular to the direc-tion of transmission. The longitudinal ultrasonic properties of a material arefundamentally a function of the compressibility of a material and theamount of material to be compressed (i.e., density), Shear waves arehighly attenuated in most fluids and so they are rarely used to characterizefood emulsions and are not discussed further here.

The parameters that are most commonly measured in an ultrasonicexperiment are the ultrasonic velocity, c, and the ultrasonic attenuationcoefficient, �:

c ¼d

t¼ f ð1Þ

A

A0¼ e��d ð2Þ


where d is the distance traveled by the wave in t, is the wavelength ofthe ultrasonic wave, f is the frequency of the ultrasonic wave, and Aand A0 are the initial and final amplitudes of the ultrasonic wave.The overall ultrasonic characteristics of a material are represented bythe complex propagation constant, k¼!/cþ i�, where ! is the angularfrequency (¼2�f ) and i is

ffiffiffiffiffiffiffi�1

p. Velocity and attenuation are funda-

mental physical properties of the emulsion but are only useful ifthey can be measured accurately and related to the droplet properties ofinterest.

During recent years, there have been considerable advances in thedevelopment of mathematical theories to describe the propagation of ultra-sonic waves through emulsions (3–5). These theories can be used to relatethe measurable ultrasonic properties of an emulsion (i.e., c, �, or k) to thedroplet characteristics of the emulsion (e.g., particle size distribution anddispersed phase volume fraction). A simplified version of the theory ispresented here, which is suitable for application to relatively diluteemulsions containing nonflocculated droplets in the long-wavelength limit(i.e., r ) (5):

K

k1

� �2

¼ 1�3i�Am

ðk1rÞ3

� �

1�9i�Ad

ðk1rÞ3

� �

, ð3Þ

where K is the complex propagation constant of the emulsion (¼!/cþ i�),k1 is the complex propagation constant of the continuous phase, � is thedispersed phase volume fraction, and Am and Ad are the monopole anddipole scattering coefficients, respectively, individual droplets. Scatteringcoefficients appropriate for fluid-in-fluid colloidal dispersions and forsolid-in-fluid colloidal dispersions are available in the literature (6).Nevertheless, calculation of these coefficients requires knowledge of manyphysiochemical properties of the oil and water phases (e.g., ultrasonicvelocity, attenuation coefficient, specific heat capacity, thermal conductiv-ity, density, and cubical expansion coefficient). The need for all of thisinformation in order to interpret ultrasonic measurements is currently oneof the major drawbacks of ultrasonic analysis techniques, although anincreasing amount of data is being tabulated in the literature (7). The ultra-sonic velocity and attenuation coefficient of an emulsion are determinedfrom the complex propagation constant using the following relationships:c¼!/Re(K) and �¼ Im(K).

It is informative to examine the physical significance of the scatteringcoefficients (Am and Ad) that appear in the theories used to model theultrasonic properties of emulsions.


1. Monopole Scattering Coefficient, Am

The droplets and surrounding liquid expand and contract to differentextents in the presence of the pressure fluctuations associated with anultrasonic wave (Fig. 1a). Consequently, the droplet pulsates relative tothe surrounding liquid, thus generating a monopole pressure wave thatpropagates into the surrounding liquid. Most of this monopole wave isnot detected in the forward direction, which leads to an increase in theattenuation of the emulsion referred to as the monopole scattering loss. Inaddition, because of pressure–temperature coupling, there is a fluctuatingtemperature gradient between the droplet and surrounding liquid thatcauses heat to flow backward and forward across the droplet interface.This process is usually irreversible because the heat flow into the dropletis less than the heat flow outward. As a result, some of the ultrasonic energyis converted to heat, which leads to an increase in the attenuation of theemulsion referred to as the thermal absorption loss. The distance thatthe thermal wave propagates into the surrounding liquid is characterizedby the thermal skin depth, T:

T ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2�1

�1CP�1!

s

ð4Þ

where the angular frequency is !, the density is �, the specific heat capacityis CP, the thermal conductivity is �, and the viscosity is � (subscript 1 refersto the continuous phase). This equation indicates that the thickness of thethermal wave increases as the frequency of the ultrasonic wave decreases.

Figure 1 Diagram (not to scale) illustrating models of ultrasonic scattering from

an emulsion droplet: (a) thermal (monopolar) scattering: (b) viscous (dipolar)

scattering.


The thermal absorption loss is much greater than the monopole scatteringloss in the long-wavelength limit.

2. Dipole Scattering Coefficient, Ad

As an ultrasonic wave passes through an emulsion, it causes the droplets tooscillate backward and forward because of the density difference betweenthem and the surrounding liquid (Fig. 1b). The movement of the dropletsrelative to the surrounding liquid leads to the generation of a dipolar pres-sure wave. Most of this dipole wave is not detected in the forward direction,which leads to an increase in the attenuation of the emulsion referred to asthe dipole scattering loss. In addition, the oscillation of the droplet isdamped because of the viscosity of the surrounding liquid, and so someof the ultrasonic energy is lost as heat, which leads to an increase in theattenuation of the emulsion referred to as the viscoinertial absorption loss.The distance that the viscoinertial wave propagates into the surroundingliquid is characterized by by viscoinertial skin depth, S:

S ¼

ffiffiffiffiffiffiffiffi2�1�t!

s

ð5Þ

This equation indicates that the thickness of the viscoinertial wavealso increases as the frequency of the ultrasonic wave decreases (8). Theviscoinertial absorption loss is much greater than the dipole scatteringloss in the long-wavelength limit.

3. Limitations of the Theory

The above theory assumes that the emulsions are dilute enough so thatmultiple-scattering effects are unimportant and that the thermal and visco-inertial waves generated by one droplet do not overlap those generated byneighboring droplets (9). Multiple-scattering and viscoinertial overlapeffects are usually quite small in food emulsions, but thermal overlap effectscan be appreciable and lead to a significant reduction in the attenuationcoefficient compared to that predicted by Eq. (3), (9). The deviationsbetween theory and experiment increase as the ultrasonic frequencydecreases, the particle size decreases, and the droplet concentrationincreases. For example, in Fig. 2, the simple theory set out in Eq. (3) predictsthat attenuation should increase with droplet concentration, whereas theexperimental observations indicate a maximum close to � ¼ 50%. Beyondthis concentration, the thermal layers of adjacent droplets begin to overlapand each individual particle is less able to convert sound energy to heat;


hence, the total attenuation is reduced. Similarly, when considering theattenuation spectra of different concentrations of fine (Fig. 3a) and coarseemulsions (Fig. 3b), the simple theory in Eq. (3) gives an adequate fitwhen the skin overlap is not significant (i.e., large, dilute particles at highfrequency).

Recently, a number of methods of incorporating thermal overlap (10)and viscoinertial overlap (11) effects into the ultrasonic scattering theoryhave been proposed and combined to produce a comprehensive core–shellscattering theory that gives a good description of the ultrasonic properties ofemulsions over a wide range of conditions (Fig. 2; the core–shell theoryalso gives a good description of the data in Fig. 3, but these lines are notshown here).

Another major limitation of the theory is that it assumes that thedroplets are randomly dispersed in the emulsion. If the droplets undergoflocculation, then there may be an appreciable deviation between the ultra-sonic scattering theory and experimental measurements because of thermaloverlap and scattering effects. In a flocculated emulsion, the droplets are inclose proximity, which leads to a greater degree of thermal overlap and a

Figure 2 Ultrasonic attenuation coefficient (at 10MHz) of an n-hexadecane oil-in-

water emulsion (r32¼ 0.2 mm) as a function of emulsion concentration. Two

theoretical lines are shown alongside the data: the Epstein–Carhart–Alegra–

Hawley (ECAH) model and a more sophisticated core–shell scattering theory

described in the text.


Figure 3 Ultrasonic attenuation spectra for different concentrations of silicone oil

in water emulsions: (a) fine (r¼ 0.1 mm) droplets and (b) coarse (r¼ 0.56 mm).

Theoretical prediction from ECAH multiple-scattering theory is shown alongside the

experimental data.


decrease in the attenuation at low ultrasonic frequencies. In addition, theeffective particles in a flocculated emulsions are larger than in a nonfloccu-lated emulsion, which leads to increased scattering and attenuation at highfrequencies (9) (Fig. 4). Recently, a theory has been developed to relate theultrasonic properties of flocculated emulsions to floc characteristics, such asconcentration, size, and internal packing (9). This theory assumes that aflocculated emulsion can be treated as a two-phase system, which consistsof spherical ‘‘particles’’ (the flocs) dispersed in a continuous phase. The flocsare treated as an ‘‘effective medium’’ whose properties depend on the size,concentration, and packing of the droplets within them. The ultrasonicproperties of a flocculated emulsion are calculated using a two-stage proce-dure. First, the thermophysical and ultrasonic properties of the effectivemedium within the flocs are calculated using a core–shell theory. Second,the ultrasonic properties of a suspension of these flocs dispersed in acontinuous phase are calculated using ultrasonic scattering theory. Sampletheoretical predictions of the attenuation spectra of flocculated emulsionsare presented in Fig. 4. These were calculated assuming varying proportionsof the droplets (diameter¼ 1 mm) were present in flocs (diameter¼ 10 mm).This theory has been shown to give good agreement with experimentalmeasurements of flocculated oil-in-water emulsions (12).

Figure 4 Theoretical prediction of the attenuation spectra of 10% silicone oil in

water emulsions flocculated to different extents. The primary particle size is 0.25 mmand the floc size is taken as 5mm, with a packing fraction of droplets inside the flocs

of 63%.


Finally, it should be mentioned that the development, refinement,and validation of theories suitable for interpreting ultrasonic spectraof emulsions and other types of colloidal suspension is currently ahighly active area of research (13–16). Therefore it is likely that furtherdevelopments will occur in the near future.

C. Ultrasonic Measurements

Many of the ultrasonic methodologies developed for fluids characterization(17–21) are suitable for food emulsions. A typical experimental setupincludes an electrical signal generator that is used to excite an ultrasonictransducer to produce an acoustic wave which, after passing through theemulsion, is detected by a second transducer. (Alternatively, a single trans-ducer and a reflector plate may sometimes be used.) These components arearranged in two main groups: pulse and resonance methods.

1. Pulse methods. A pulse of ultrasound is measured after transmis-sion through a known path length of emulsion. This ultrasonicvelocity is determined by measuring the time taken for the pulse totravel a fixed distance [Eq. (1)], and the attenuation coefficient isdetermined by measuring the reduction in amplitude of the pulseafter it has traveled through the emulsion [Eq. (2)] (17,22). Thefrequency dependence of the ultrasonic properties of an emulsioncan be determined using two different pulsed techniques: toneburst and Fourier transform. In the tone-burst method, a pulsecontaining a number of cycles of ultrasound at a single frequency(actually a narrow range of frequencies) is applied to the sample,and the ultrasonic velocity and attenuation coefficient of thesample are determined at this specific frequency. This procedure isthen repeated using ultrasonic waves with different frequencies. Inthe Fourier transform method, a broadband pulse of ultrasound,which contains a range of different frequency components, isapplied to the sample. Fourier transformation is then used todetermine the velocity and attenuation of the ultrasonic wave overthe frequency range of the transducer. The major advantage of theFourier transform method is that a spectrum can be obtainedusing a single ultrasonic pulse, which speeds up the analysis (22).

2. Resonance Methods. Analytical instruments based on theresonator method use a continuous ultrasonic wave, rather thanan ultrasonic pulse, to determine the ultrasonic properties of asample. One transducer measures signal intensity and the othergenerates a continuous wave of either constant wavelength while


the path length is varied or variable frequency across a fixed pathlength. In a fixed-path-length resonator, a continuous wavecontaining a single, slowly increasing ultrasonic frequency istransmitted across the measurement cell. When the cell pathlength is an integer number of whole wavelengths, constructiveinterference occurs and there is a maximum in the detected energy(20,21,23). The shape and position of these resonance peaks can beused to calculate the ultrasonic velocity and attenuation coeffi-cient of the liquid in the cell to very high precision.

Traditionally, analytical instruments for characterizing emulsion propertieshad to be constructed by researchers’, however, a number of instrumentmanufacturers now supply analytical instruments based on ultrasoundthat are specifically designed for characterizing emulsion properties. Thereare a number of important practical considerations that should be takeninto account when developing or selecting an analytical instrument forcharacterizing emulsion properties.

. The ultrasonic properties of most liquids are strongly temperaturedependent and so it is important to use an instrument in which thetemperature is precisely controlled (i.e., 0.2�C or better).

. The measurement volume of commercial ultrasonic analyticalinstruments can vary from a few milliliters to several liters, whichmay be important if a sample is expensive or in limited supply.

. The measurement time may be important in systems where rapidkinetic measurements are required. Data acquisition is usuallymuch faster for Fourier transform spectrometers than for ‘tone-burst’ or resonance spectrometers, enabling faster kinetic processesto be monitored.

. Many commercial spectrometers will automatically calculateparticle size distribution and disperse the phase volume fractionof an emulsion from its measured ultrasonic spectrum. The theoriesused by commercial instruments to make this calculation varywidely and it is essential to ensure that the theory has beenvalidated for the types of system under investigation.

. Whereas both velocity and attenuation can be readily measured,velocity tends to be more affected by imprecision in temperatureand composition, so attenuation is often more useful for structuralmeasurements.

. Probably the single most common reason for poor-qualityultrasonic measurements is the presence of small air bubblestrapped within liquids. The large impedance mismatch between aircells and the surrounding emulsion leads to extensive scattering of


ultrasound that can obscure the effects of the emulsion itself. Inmany cases, air cells can be eliminated by judicious use of gentlecentrifugation or degassing prior to measurement.

II. APPLICATIONS

Studies over many years, usually using custom-built ultrasonic instrumenta-tion, has demonstrated the potential of the ultrasonic technique for char-acterizing the properties of emulsion droplets (e.g., droplet size, dropletconcentration, droplet crystallization, and creaming stability). Ultrasoundhas advantages over many alternative technologies because measurementsare rapid, nondestructive, and noninvasive and can be made in concentratedor optically opaque materials. Recently, analytical instruments based onultrasonic spectrometry that are specifically designed for characterizingemulsion properties have become commercially available. Consequently,many of the potential benefits of ultrasonic technology for characterizingfood emulsions have become available to scientists working in the foodindustry. Nevertheless, it is important for users of these commercial instru-ments to be aware of the theoretical basis of the ultrasonic technique,because improper measurement procedures or data interpretation can leadto appreciable errors.

A. Droplet Concentration

Analysis of droplet concentration is one of the simplest and most successfulapplications of low-intensity ultrasonic techniques to food emulsions. Thisapplication is based on there being an appreciable difference in the ultra-sonic velocities of the oil and aqueous phases when ultrasonic velocity mea-surements are used or there being a significant amount of attenuation of theultrasonic wave due to the presence of the emulsion droplets when attenua-tion coefficient measurements are used. The influence of droplet concentra-tion on the ultrasonic velocity and attenuation of a model food emulsion isshown in Fig. 2. The droplet concentration of an emulsion can be deter-mined by developing an empirical calibration curve or by using an appro-priate theoretical model. The ultrasonic scattering theory described earlier[ECAH theory, Eq. (3)] adequately predicts the effects of droplet concentra-tion on the ultrasonic properties of emulsions, provided that thermal over-lap effects are not appreciable (24,25) (i.e., � < 10% in this case), but whenthermal overlap becomes important (small droplet sizes, low ultrasonic fre-quencies, or high droplet concentrations), a more sophisticated theory mustbe applied (24). Criteria for deciding whether these effect should be taken


into account have been published in the literature (24). Ultrasonic measure-ments have previously been used to measure the droplet volume fraction ofemulsified vegetable oil (26), milk (27), and salad dressings (28, 29).

B. Particle Size

Analytical instruments based on ultrasonic spectroscopy are gaining increas-ing acceptance for measuring the droplet size distribution of concentratedfood emulsions. These instruments normally measure the ultrasonic attenua-tion coefficient (and sometimes the ultrasonic velocity) over a range offrequencies and then compute the particle size distribution (and volumefraction) that gives the best agreement between the measured spectra andthat predicted by ultrasonic scattering theory. The suitability of a giveninstrument for a particular application largely depends on how accuratelythe scattering theory it uses to analyze the measured spectrum represents theultrasonic properties of the emulsion being analyzed. In relatively dilutenonflocculated emulsions, there is excellent agreement between theory andexperiment, and the measured particle size distributions are in close agree-ment with those measured by other technologies (24,25). However, whenthermal overlap effects are important (e.g., in concentrated or flocculatedemulsions), there may be appreciable differences between theory and experi-ment (Figs. 2 and 3), which would lead to large erros in any measuredparticle size distributions. For example, if the ECAH theory was used tocharacterize the concentrated emulsion in Fig. 3a, it would attempt to cal-culate the particle size (r) in Eq. (3) that would cause the theoretical line tobest approach the measured data. However, under these conditions, theECAH theory does not well describe the emulsion (particularly at lowfrequencies), so the particle size calculated would be meaningless.Therefore, one should always ensure that the ultrasonic scattering theorybeing used to interpret the data is appropriate for all sizes, concentrations,and frequencies (24).

One should also be aware that in order to compute the droplet char-acteristics from the measured ultrasonic spectrum in a reasonable time, it isoften necessary to make some simplifying assumption about the shape of theparticle size distribution (e.g., it is monomodal or bimodal or has a log-normal form). Only if these assumptions about the shape of the particle sizedistribution are incorrect will the final data be inaccurate. Despite theselimitations, ultrasonic spectroscopy is currently one of the few practicalmethods of measuring the droplet size distribution of concentrated emul-sions in situ [other techniques include nuclear magnetic resonance (NMR),electroacoustics, diffuse optical reflectance, and dielectric spectroscopy] (1).The ultrasonic technique has been used to determine particle size


distributions in a number of food products, including emulsified vegetableoils (30), salad cream (29), and milk fat globules (27).

C. Droplet Flocculation

The bulk physicochemical properties of many food emulsions are stronglyinfluenced by droplet flocculation (e.g., creaming and rhelogy) (1) . It istherefore important to have analytical techniques capable of providinginformation about the extent of droplet flocculation and about thestructural characteristics of the flocs formed. Characterization of floc prop-erties in most food emulsions is extremely difficult because their delicatestructure is disrupted by the sample preparation procedures required bymany analytical techniques (e.g., dilution or stirring in light scattering, elec-trical pulse counting, or microscopy techniques). A number of studies haveshown that the ultrasonic properties of emulsions change appreciably whenthe droplets become flocculated (8,31–33) and ultrasonic attenuationspectroscopy has been used to study droplet flocculation in oil-in-wateremulsions induced by diverse mechanisms (12,34,35) in the light of recenttheoretical developments (9).

The same ultrasonic spectroscopy technique has been used to study thedisruption of flocs in a shear field (36). As the emulsions are exposed tohigher shear rates, the flocs become disrupted and their attenuation spectrabecomes closer to that of nonflocculated droplets. The sensitivity of ultra-sonic spectra to droplet flocculation and the fact that ultrasonic techniquescan be applied nondestructively and noninvasively opens up the potentialfor ‘‘rheo-acoustic’’ techniques (37,38). These techniques measure the rheol-ogy of colloidal dispersions using a conventional rheometer while obtaininginformation about the structure and concentration of flocs using an ultra-sonic spectrometer. Thus, it is possible to correlate changes in dispersionrheology to changes in floc structure as the shear rate applied to the systemis varied in a controlled manner.

D. Gravitational Separation

Creaming leads to a change in oil distribution throughout a product. Ultra-sonic velocity (or attenuation) can be used to measure the concentration ofoil in an emulsion (see Sect. II.A) as a function of height using a simple one-dimensional imaging apparatus. A stepper motor moves an ultrasonictransducer along the vertical axis of the emulsion and a series of ultrasonicmeasurements are made at different heights and times (39). The ultra-sonic measurements can then be converted into oil concentrations using acalibration curve or ultrasonic scattering theory. A typical creaming profile


for emulsified corn oil is shown in Fig. 5. In this case, the ultrasonic data aregiven as a two-dimensional image taken across a sample of emulsion atdifferent times after preparation. Light areas indicate a higher oil concen-tration. This type of approach often has limited spatial resolution becausethe sound beam has a significant width (often �1 cm) and averages theultrasonic properties over that distance. It would be possible to achievebetter resolution with a focused transducer or more sophisticated signalprocessing. It is important to remember that converting an ultrasonic mea-surement to oil concentration is not straightforward, as both particle sizeand oil concentration can vary within a cream layer and both can affect theultrasonic signal (40–43). Even with these limitations, the use of ultrasonicsto probe the structure of cream layers is a powerful tool and one of theunique applications of the technique. Conventional measurements of cream-ing are often based on visual identification of the boundary between and oil-rich layer and oil-poor layer and can easily miss details of fine structure. Forexample, xanthan gum can cause an emulsion to cream due to depletionflocculation (8), but there are a variety of structures formed within thecream layer not evident to the naked eye. Ultrasonic imaging of the processwas also able to detect the influence of added salt on the structure of thecream layer induced in the emulsion by xanthan gum (44).

E. Droplet Crystallization

Ultrasound has proved to be a highly sensitive method for monitoring themelting and crystallization for emulsion droplets (45,46). The physicochem-ical basis of the utilization of ultrasound for this purpose is the relatively

Figure 5 Diagrams and ultrasonic images illustrating creaming in oil-in-water

emulsions. In the ultrasonic images, light regions indicate low oil content and dark

regions indicate high oil content. Initially, the droplets are evenly dispersed and the

ultrasonic image shows no detail. After storage, the droplets accumulate at the top of

the container and a clear light/dark band is seen in the ultrasonic image.


large difference in physical and ultrasonic properties of solid and liquiddroplets.

The influence of heating and cooling on the ultrasonic velocity andattenuation coefficient of an n-hexadecane oil-in-water emulsion is shown inFig. 6. When an emulsion containing crystalline fat droplets is heated, thereis a sharp decrease in the ultrasonic velocity around the melting point of thebulk fat because the solid phase has a higher ultrasonic velocity than theliquid phase. When the same emulsion containing liquid oil droplets iscooled, the droplets exhibit a high degree of supercooling before they crys-tallize and the ultrasonic velocity returns to the starting solid-fat line. Theemulsion droplets do not crystallize until they are cooled well below theirthermodynamic melting point because the impurities that normally promoteheterogeneous nucleation in bulk oils are distributed throughout such a verylarge number of droplets that each individual droplet is free of impurities.Consequently, nucleation tends to take place by a homogeneous mechanism(or sometimes a surface heterogeneous mechanism) in emulsions rather thanbe promoted by impurities as in bulk oils (47). To a first approximation,the fraction of crystalline material within the droplets (�) at a particulartemperature can then be calculated as

� ¼1

c2�

1

c2l

� � 1

c2s�

1

c2l

� �

ð6Þ

Figure 6 The effect of a cooling–heating cycle on the ultrasonic velocity in a 20%

hexadecane-in-water emulsion. The speed of sound in the emulsion decreases with

temperature and there is an abrupt change corresponding to the phase transition in

the droplet oil. Supercooling of the liquid oil is responsible for the hysteresis loop

observed.


where c is the measured ultrasonic velocity of the emulsion and cs and cl arethe ultrasonic velocities in the emulsion when the droplets are completelysolidified or completely liquefied oil, respectively, extrapolated to the mea-surement temperature (48). Equation (6) is based on the assumption thatsolid fats and liquid oils have similar densities and behave ideally as amixture (49). Ultrasonic methods have been successfully used to measurethe supercooling of emulsified triacyl glycerols (45,48), margarine andbutter (50), and milk fat (49).

A similar hysteresis loop occurs in the attenuation coefficient of anemulsion during the melting and freezing of the droplets (51). However, inthe experiment reported in Fig. 6 there was a large attenuation peak andextensive velocity dispersion at the melting point of the dispersed phase. Thisoccurs because pressure–temperature fluctuations associated with the acous-tic wave perturb the solid–liquid equilibrium and some of the ultrasonicenergy is lost as heat. The magnitude of the attenuation peak is dependenton the frequency of the ultrasound. At low frequencies, the temperature–pressure gradients associated with the ultrasonic wave are small and so theequilibrium is maintained and attenuation is low. At high frequencies, thechanges in temperature and pressure occur so rapidly that there is no time forthe system to respond and the ultrasonic absorption is also low. However, atintermediate frequencies, the temperature–pressure gradients are sufficientlylarge and persist for a sufficiently long time that an appreciable amount ofmaterial is able to undergo an ultrasonically induced phase transition. Theenergy required to rapidly crystallize and melt a proportion of the fat leads tothe significant absorption of the ultrasonic wave observed. The excessabsorption is related to the dynamics of the phase transition (52).

III. CONCLUSIONS

The ultrasonic properties of emulsions are relatively easy to measure togood precision, particularly considering the wide commercialisation ofacoustic spectrometers suitable for fluids characterization. The developmentof off-the-shelf instrumentation has been accompanied by advances in thetheory describing ultrasonic spectra of emulsions in terms of particle size,concentration, association, and distribution. By comparing theory andexperiment, it is possible to measure dispersed phase volume fraction,particle size (and, to some extent, particle size distribution), degree offlocculation, and creaming. Although examples of these measurementshave been reported in the literature, rarely have multiple unknowns beendeconvoluted from a single acoustic spectrum and so comprehensive ultra-sonic characterization of droplet properties remains elusive. It is probably


more practical to use ultrasound in conjunction with other established meth-ods that can measure other properties (e.g., density measurements to calcu-late dispersed phase volume fraction and light scattering to calculateprimary particle size) and then use the acoustic data to measure emulsionproperties not otherwise readily accessible (e.g., flocculation).

Very often our understanding of the world jumps forward when wedevelop a new way of looking at it. As an example, consider the widespreadintroduction of static light-scattering instruments into food emulsionslaboratories over the past 20 years. The theories and phenomena importantin food emulsion science did not change, but the explanations used began tomore specifically describe events in terms of the size of the structures pre-sent. In this example, the principles of light scattering had a long history, butthe difficulty of making measurements and the complexity of the mathe-matics prevented widespread application until they were integrated into aconvenient package.

Ultrasonic instruments for particle characterization are in a similarposition now as light-scattering instruments were 20 years ago and we canexpect a similar pattern of development. Using these methods, we can rea-sonably measure realistic concentrations of fat in a relatively undisturbedstate and, hence, evaluate the real properties of delicate structures that ourexisting techniques necessarily disrupt in sample preparation. It is our hopethat this chapter will go some way to encouraging food emulsion scientiststo gain the benefits of the new approaches while avoiding some of the pit-falls of poor measurement technique or extension of a theory beyond itsassumptions.

REFERENCES

1. D. J. McClements, Food Emulsions: Principles, Practice, and Techniques, CRC

Press, Boca Raton, FL, 1999.

2. M. J. W. Povey, Ultrasonic Techniques for Fluids Characterization, Academic

Press, San Diego, CA, 1997.

3. A. S. Dukhin, P. J. Goetz, T. H. Wines, and P. Somasundaran, Acoustic and

electroacoustic spectroscopy, Colloids Surfaces A 173, 127–158 (2000).

4. D. J. McClements. Principles of ultrasonic droplet size determination in

emulsions, Langmuir 12, 3454–3461 (1996)

5. D. J. McClements and J. N. Coupland, Theory of droplet size distribution

measurements in emulsions using ultrasonic spectroscopy, Colloids Surfaces A

117, 161–170 (1996).

6. J. A. Allegra and S. A. Hawley, Attenuation of sound in suspensions

and emulsions: Theory and experiments, J. Acoust. Soc. Am. 51, 1545–1564

(1971).


7. J. N. Coupland and D. J. McClements, Physical properties of liquid edible oils,

J. Am. Oil Chem. Soc. 74, 1559–1564 (1997).

8. D. J. McClements Ultrasonic determination of depletion flocculation in oil-in-

water emulsions containing a non-ionic surfactant, Colloids Surfaces A 90,

25–35 (1994).

9. D. J. McClements, W. Herrmann, and Y. Hemar, Influence of flocculation on

the ultrasonic properties of emulsions: theory, J. Phys. D 31, 2950–2955 (1998).

10. D. J. McClements, Y. Hemar, and N. Herrmann, Incorporation of thermal

overlap effects into multiple scattering theory, J. Acoust. Soc. Am., 105,

915–918 (1999)

11. A. Dukhin and P. Goetz, Acoustic and electroacoustic spectroscopy

characterizing concentrated dispersions emulsions, Adv. Colloid Interf. Sci.

92, 73–132 (2001).

12. R. Chanamai, N. Herrman, and D. J. McClements, The influence of

flocculation on the ultrasonic properties of emulsions: J. Phys. D 31,

2956–2963 (1998).

13. O. G. Harlen, M. J. Holmes, M. J. W. Povey, Y. Qiu, and B. D. Sleeman, A low

frequency potential scattering description of acoustic propagation in

dispersions, SIAM J. Appl. Math. 61, 1906–1931 (2001).

14. P. D. M. Spelt, M. A. Norato, A. S. Sangani, and L. L. Tavlardies.

Determination of particle size distributions from acoustic wave propagation

measurements, Phys. Fluids 11, 1065–1080, (1999).

15. P. D. M. Spelt, M. A. Norato, A. S. Sangani, M. S. Greenwood, and L. L.

Tavlarides, Attenuation of sound in concentrated suspensions: theory and

experiments, J. Fluid Mech. 430, 51–86 (2001).

16. F. Vander Meulen, G. Feuillard, O. B. Matar, F. Levassort, and M. Lethieco,

Theoretical and experimental study of the influence of the particle size

distribution on acoustic wave properties of strongly inhomogeneous media,

J. Acoust. Soc. Am. 110, 2301–2307 (2001).

17. D. J. McClements and P. Fairley, Frequency scanning ultrasonic pulse echo

reflectometer, Ultrasonics 30, 403–405, (1992).

18. E. P. Papadakis, The measurement of ultrasonic velocity, Ultrasonic

Measurement Methods (R.N. Thurston and A.D. Pierce, eds.), Academic

Press, San Diego, CA, 1990, pp. 81–106.

19. E. P. Papadakis, The measurement of ultrasonic velocity, Ultrasonic

Measurement Methods (R.N. Thurston and A.D. Pierce, eds.), Academic

Press, San Diego, CA, 1990, pp.107–155.

20. A.P. Sarvazyan, Development of methods of precise ultrasonic measurements

in small volumes of liquids, Ultrasonics 20, 151–154 (1982).

21. A. P. Sarvazyan and T. V. Chalikian, Theoretical analysis of an ultrasonic

interferometer for precise measurements at high pressures, Ultrasonics 29,

119–124, (1991).

22. D. J. McClements and P. Fairley, Ultrasonic pulse echo reflectometer,

Ultrasonics 29, 58–62, (1991).

23. J. Blitz, Ultrasonics: Methods and Applications, Butterworths, London, 1971.


24. R. Chanamai, N. Herrman, D. J. McClements, Influence of thermal overlap

effects on the ultrasonic attenuation spectra of polydisperse oil-in-water


25. N. Herrmann and D. J. McClements, Ultrasonic propagation in highly

concentrated oil-in-water emulsions, Langmuir 15, 7937–7939 (1999).

26. D. J. McClements and M. J. W. Povey, Scattering of ultrasound by emulsions,

J. Phys. D: Appl. Phys. 22, 38–47 (1989).

27. C. A. Miles, D. Shore, and K. R. Langley, Attenuation of ultrasound in milks

and creams, Ultrasonics 28, 394–400 (1990).

28. R. Chanamai, F. Alba, and D. J. McClements, Ultrasonic spectroscopy study

of salad dressings, J. Food Sci. 65, 507–513 (2000).

29. D. J. McClements, M. J. W. Povey, M. Jury, and E. Betsanis, Ultrasonic

characterization of a food emulsion, Ultrasonics 28, 266–272 (1990).

30. J. N. Coupland and D. J. McClements, Droplet size determination in food

emulsions: comparison of ultrasonic and light scattering methods, J. Food Eng.

50, 117–120, (2001).

31. Y. Hermar, N. Hermann, P. Lemarechal, R. Hocquart, and F. Lequeux,

Effective medium model for ultrasonic attenuation due to the thermo-elastic

effect in concentrated emulsions, J. Phys. II 7, 637–647 (1997).

32. D. Hibberd, A. Holmes, M. Garrood, A. Fillery-Travis, M. Robbins, and R.

Challis, Ultrasonic monitoring of oil-in-water emulsions undergoing depletion

flocculation, J. Colloid Interf. Sci. 193, 77–87, (1997).

33. D. J. Hibberd, B. H. Robinson, and M. M. Robins, Ultrasonic characterisation

of colloidal dispersions: detection of flocculation and adsorbed layers, Colloids

Surfaces B 12, 359–371 (1999)

34. R. Chanamai, N. Herrmann, and D. J. McClements, Probing floc structure by

ultrasonic spectroscopy, viscometry, and creaming measurements, Langmuir

16, 5884–5891 (2000).

35. N. Herrmann, Y. Hemar, P. Lemarechal, and D. J. McClements, Probing

particle–particle interactions in flocculated oil-in-water emulsions using

ultrasonic attenuation spectrometry, Eur. Phys. J. E 5, 183–188 (2001).

36. R. Chanamai, N. Herrmann, and D. J. McClements, Ultrasonic spectroscopy

study of flocculation and shear-induced floc disruption in oil-in-water

emulsions. J. Colloid Interf. Sci. 204, 268–276 (1998).

37. L. Haider, P. Snabre, and M. Boynard, Rheo-acoustical study of the shear

disruption of reversible aggregates. Ultrasound scattering from concentrated

suspensions of red cell aggregates, J. Acoust. Soc. Am. 107, 1715–1726 (2000).

38. P. Snabre, L. Haider, and M. Boynard, Ultrasound and light scattering from a

suspension of reversible fractal clusters in shear flow, Eur. Phys. J. E. 1, 41–53

(2000).

39. T. K. Basaran, K. Demetriades, and D. J. McClements, Ultrasonic imaging of

gravitational separation in emulsions, Colloids Surfaces A 136, 169–181 (1998).

40. V. J. Pinfield, E. Dickinson, and M. J. W. Povey, Modeling of concentration

profiles and ultrasonic velocity profiles in a creaming emulsion: importance of

scattering effects, J. Colloid Interf. Sci. 166, 363–374 (1994).


41. V. J. Pinfield, M. J. W. Povey, and E. Dickinson, The application of modified

forms of the Uric equation to the interpretation of ultrasonic velocity in

scattering systems, Ultrasonics 33, 243–251 (1995).

42. V. J. Pinfield, M. J. W. Povey, and E. Dickinson, Interpretation of ultrasonic

sound velocity in creaming profiles, Ultrasonics 34, 695–698 (1996).

43. V. J. Pinfield, M. J. W. Povey, and E. Dickinson, Modeling of combined

creaming and flocculation in emulsions, J. Colloid Interf. Sci. 186, 80–89 (1997).

44. E. Dickinson, J. Ma, and M. J. W. Povey, Creaming of concentrated oil in

water emulsions containing xanthan, Food Hydrocolloids 8, 481–497 (1994).

45. W. Kloek, P. Walstra, and T. Van Vliet, Nucleation kinetics of emulsified

triglyceride mixtures, J. Am. Oil Chem. Soc. 77, 643–652 (2000).

46. D. J. McClements, E. Dickinson, and M. J. W. Povey, Crystallization of

hydrocarbon-in-water emulsions containing a mixture of solid and liquid

droplets, Chem. Phys. Lett. 172, 449, (1990).

47. D. J. McClements, S. R. Dungan, J. B. German, C. Simoneau, and J. E.

Kinsella, Droplet size and emulsifier type affect crystallization and melting of

hydrocarbon-in-water emulsions, J. Food Sci. 58, 1148 (1993).

48. J. N. Coupland, E. Dickinson, D. J. McClements, M. J. W. Povey, and C.

Rancourt de Mimmerand, Crystallization in simple paraffins and monoacid

saturated triacylglycerols dispersed in water, (E. Dickinson and P. Walstra,

eds.), Food Colloids and Polymers: Stability and Mechanical Properties, Royal

Society of Chemistry, London, 1993.

49. C. A. Miles, G. A. J. Fursey, and R. C. D. Jones, Ultrasonic estimation of

solid/liqud ratios in fats, oils, and adipose tissue, J. Sci. Food Agric. 36, 218–228

(1985).

50. D. J. McClements, The use of ultrasonics for characterizing fats and emulsions,

Ph.D. dissertation, Leeds University, 1988.

51. D. J. McClements, M. J. W. Povey, and E. Dickinson, Absorption and velocity

dispersion due to crystallization and melting of emulsion droplets, Ultrasonics

31, 433–437 (1993).

52. A. J. Matheson, Molecular Acoustics, Wiley, London, 1971.


15NMR Studies of Emulsionswith Particular Emphasison Food Emulsions

Balin BalinovAmersham Health, Oslo, Norway

Francois MarietteCemagref, Rennes, France

Olle SodermanUniversity of Lund, Lund, Sweden

I. INTRODUCTION

Very early in the development of colloid science, emulsions received con-siderable attention. This is due to the fact that emulsions are of great funda-mental as well as technical importance. They occur in a multitude ofsituations ranging from the extraction of crude oil to various food products.Therefore, it is not surprising that the practical knowledge about emulsionsis quite extensive. People engaged in the production generally know how toproduce emulsions with desired properties such as droplet-size distributionsand shelf-life. The same level of empirical knowledge is at hand for theopposite process of breaking an emulsion. When it comes to the basic scien-tific understanding of emulsions, we are a little worse off. It is quite clearthat several fundamental properties of emulsions, such as what factorsdetermine the stability of emulsions, what is the importance of the proper-ties of the continuous phase, and so on, are not fully understood. To someextent, we also lack or have not yet applied suitable techniques in the studyof emulsions.


The majority of emulsions are stabilized by surfactants and Bancroft,one of the pioneers in emulsion science, realized that the stability of anemulsion was related to the properties of the surfactant film. This insighthas become important during the latter years, when attempts have been madeto draw on the rather detailed and profound understanding that today existsabout bulk surfactant systems in the description of emulsions (1). This highlevel of understanding about bulk surfactant systems stems not the least fromthe application of modern physicochemical techniques such as scatteringmethods and nuclear magnetic resonance (NMR). It seems reasonable toexpect that the application of these techniques to emulsion systems wouldlead to an increased basic understanding of central emulsion properties.

In this contribution, we will attempt to show how NMR can be used tostudy various emulsion systems and how NMR may be used in emulsionapplications. We focus in particular on food emulsions, which, by their verynature, cover a very wide area in practical applications. One finds ‘‘semi-solid’’ varieties such as margarine and butter as well as liquid ones such asmilk, sauces, dressings, and various beverages. In addition, food emulsionsalso include an array of products that contain both solids and/or gases inaddition to two liquid phases such as in ice cream. Because of the complexityin their chemical composition and their heterogeneous structure, the use ofNMR techniques in the study of food emulsions are up until now relativelylimited. This is in contrast to the many studies dealing with biopolymers andtheir structure modification induced by food processing. The first applica-tions of NMR to food emulsions were in the determination of solid-fatcontent and emulsion composition. These methods are based on the relativeNMR signal intensities. Somewhat surprisingly, the relaxation time varia-tion as a function of fat and/or water phase structure and composition hasreceived little attention. Since the development of new benchtop NMRspectrometers, new applications using both relaxation time parametersand self-diffusion coefficient have been suggested and introduced.

We first discuss the NMR technique as such, with special emphasizeon features important for the study of emulsions. Subsequently, we treatsome important examples of how NMR can be used to obtain importantinformation about central aspects of emulsions, in general, and food emul-sions, in particular.

II. THE NMR TECHNIQUE

A. Fundamentals

Nuclear magnetic resonance is a spectroscopic technique and as such it isvery well suited for investigations of topics such as molecular arrangements


and molecular dynamics. NMR has gone through a dramatic developmentover the last decades. For this reason, there are numerous monographstreating many different aspects of the method (2–4). Here, we will attemptto introduce the technique from the point of view of emulsion science.

Nuclear magnetic resonance spectroscopy is based on the fact thatsome nuclei possess a permanent nuclear magnetic moment. When placedin an external magnetic field, they take a certain well-defined state, whichcorrespond to distinct energy levels. Transitions between neighboring energylevels take place due to absorption of electromagnetic radiation at radio-frequencies. In NMR, the signal is obtained by a simultaneous excitation ofall transitions with radio-frequency pulses, followed by detection of theirradiation emitted as the system returns to the equilibrium state. Therecorded free-induction decay (FID) may be used as such, but Fouriertransformation of the time-resolved FID is usually performed to obtainthe NMR spectrum.

Modern NMR spectrometers, operating at 500–800 MHz for protons,have a high resolution that allows one to identify up to several hundreds oflines in a complex NMR spectrum. In most cases, only a limited number oflines are used to obtain the information needed. In some applications, thespectral resolution is not a necessary step in the data analyses and theinformation may be obtained from the time dependence of the amplitudeof the FID. The FID reflects all of the nuclei of a given type (e.g., protons)and is used to obtain information on the average relaxation phenomena ofthe nuclei of interest. For this type of signal detection, 10–20-MHz NMRspectrometers for protons is an option suitable for many industrial applica-tions.

Nuclear magnetic resonance is a very versatile spectroscopic techniquefor three main reasons. (1) It is not a destructive technique. Thus, the systemmay be studied without any perturbation that influences the outcome andinterpretations of the measurements. The system can be characterizedrepeatedly with no time-consuming sample preparations in between runs.(2) There are a large number of spectroscopic parameters that may bedetermined by NMR relating to both static and dynamic aspects of awide variety of systems. (3) A large number of atomic nuclei carry nuclearspins, and for essentially any element in the periodic table, it is possible tofind at least one suitable nucleus. This has the important consequence thatfor systems that show local segregation (such as surfactant and emulsionsystems), it is possible to investigate different domains of the microhetero-geneous system by studying different nuclei.

The most commonly used nucleus in NMR is 1H, but also 13C, 19F,23Na and 31P are of importance. These nuclei are naturally occurringisotopes and need not be inserted chemically. Another very useful nucleus


is 2H; however, the natural abundance of this species is, in general, too smallto allow for reasonable measuring times. Therefore, this nucleus is usuallyinserted by chemical labeling. This can, in fact, be used to an advantagebecause the 2H nucleus can be directed to a particular part of the molecule,hence making it possible to investigate different parts of a molecule.

It is convenient to divide the parameters obtained from NMR spectrainto dynamic and static parameters.

B. Static Parameters

The static parameters are obtained from the observed resonance frequencies.The general frequency range where a particular nucleus shows a spectro-scopic line is determined by the magnetogyric ratio, which is a nuclearproperty without chemical interest. However, the precise value of the reso-nance frequency is determined by molecular properties. For isotropic sys-tems, the two most important parameters determining the resonancefrequency is the chemical shift and the scalar spin–spin coupling.

The chemical shift is determined by the screening due to the electronsin the vicinity of the investigated nucleus. This, in turn, is determined mainlyby the primary chemical structure of the molecule but also other factorssuch as hydrogen-bonding, conformation, and the polarity of the environ-ment influence the chemical shift. In systems with local segregation, such asemulsions, this has the consequence that one may observe two separatesignals from the same molecule if it resides in two different environments.This occurs if the exchange of the molecule between the two environments isslow on the relevant timescale, which is given by the inverse of the shiftdifference between the two environments. This feature of the experimentgives us the possibility to investigate the distribution of molecules andexchange rates in microheterogeneous systems.

The scalar coupling is exclusively of intramolecular origin and of littleimportance in emulsion systems.

Another static parameter of importance is the integrated area under agiven peak. This quantity is proportional to the number of spins contribut-ing to a given peak, and hence (relative) concentrations can be determined.This, as will be shown in Sect. VI, has some important applications inemulsion science.

C. Dynamic Parameters

Dynamic processes on the molecular level influence the nuclear spin systemby rendering the spin Hamiltonian time dependent. Depending on the rela-tion between the characteristic times of the molecular motions, �c, and the


strength of the modulated interaction, !I (expressed in frequency units), onecan identify different regimes.

For slow motions, when �c!I� 1, the system is in the solid regime and!I contributes to the resonance frequencies observed. This is the situationencountered in anisotropic liquid crystals, where the NMR spectra areusually dominated by static effects.

In the other extreme, when �c!I 1, the phenomenon of motionalnarrowing occurs. Here, the spin relaxation is characterized by a limitednumber of time constants, the most readily observed are the longitudinal,T1, and transverse, T2, relaxation times. The first of these characterizes thedecay of the Mz magnetization to equilibrium and the second characterizesthe return of the Mx,y magnetization to equilibrium (where Mz is perpendi-cular to the polarizing magnetic field and Mx,y lies in a plane perpendicularto Mz). Measurements of the relaxation times require that the sample isperturbed by various pulse sequences. The one most often used for measur-ing T1 is the inversion recovery sequence (or variations of it), and differentecho sequences are usually used for measuring T2.

For the case when �c!I� 1, the shape of the NMR spectral line isstrongly affected by the characteristic features of the molecular motions.

We shall make two remarks concerning features, which are peculiar tothe topic of NMR and emulsions. The first deals with the fact that anemulsion system may actually contain molecules that fulfill more than oneof the motional regimes described earlier. Consider, for instance, moleculesin the continuous phase. For them, the condition �c!I 1 holds. Formolecules residing at the interface between the two liquids of the emulsion,the situation may be different, and depending on the size of the emulsiondroplet, they may experience any of the three time domains describedearlier.

Our second remark deals with the effect of diamagnetic susceptibilityvariations. In heterogeneous structures, the magnetic field is perturbed inthe vicinity of regions of differing susceptibilities. As a consequence, therandom motions of the molecules in this spatially varying magnetic fieldinduces relaxation effects. T1 and T2 relaxation require fluctuations at theLarmor frequency, whereas, in addition, T2 relaxation is also sensitive toslower fluctuations. Because the fluctuations brought about by diffusion inthe locally varying field typically occur with rates, which are slower than theLarmor frequencies, T2 T1 for this effect. Peculiar to systems with sphe-rical structures, such as reasonably dilute emulsions, is that the gradient ofthe magnetic field inside the droplet is not affected. Thus, the above-described effect only operates for the continuous phase, which can beused to identify whether an emulsion is of the oil-in-water or water-in-oiltype in a straightforward manner (5).


D. Measurement of Self-Diffusion

A very important dynamic molecular process is that of the transport ofmolecules due to thermal motion. This can conveniently be followed bythe NMR pulsed field gradient (PFG) method. Because this approach hasbeen of particular importance in the field of microheterogenous surfactantsystems, in general, and in emulsion systems, in particular, we will spendsome time introducing this application of NMR. The technique has recentlybeen described in a number of review articles (cf. Refs. 2 and 6–8), so herewe will merely state that the technique requires no addition of probes, thusavoiding possible disturbances that addition of probes may cause, and thatit gives component resolved self-diffusion coefficients with great precision ina minimum of measuring time. The main nucleus studied is the proton, butother nuclei, such as Li, F, Cs, and P, are also of interest.

The method monitors transport over macroscopic distances (typicallyin the micrometer regime). Therefore, when the method is used to studycolloidal systems, the determined diffusion coefficients reflect aggregatesizes and obstruction effects for colloidal particles. This is the origin ofthe success the method has had in the study of microstructures of surfac-tant solutions and also forms the basis of its applications to emulsionsystems. We expect that the PFG method will also be increasingly impor-tant in the study of emulsion systems and, therefore, we will discussthe method in some detail, with particular focus on its application toemulsions.

The fact that the information is obtained without the need to invokecomplicated models, as is the case for the NMR relaxation approach, isparticularly important. In this context, it should be stressed that thePFG approach measures the self-diffusion rather than the collectivediffusion coefficient, which is measured by, for instance, light-scatteringmethods.

In its simplest version, the method consists of two equal and rectan-gular gradient pulses of magnitude g and length , one on either side of the180� radio-frequency(rf)-pulse in a simple Hahn echo experiment. For mole-cules undergoing free (Gaussian) diffusion characterized by a single diffu-sion coefficient of magnitude D, the echo attenuation due to diffusion isgiven by (9,10)

Eð�, , gÞ ¼ E0 exp ��2g22 ��

3

� �

D

� �

ð1Þ

where � represents the distance between the leading edges of the two gra-dient pulses, � is the magnetogyric ratio of the monitored spin, and E0


denotes the echo intensity in the absence of any field gradient. By varyingeither g, , or � (while at the same time keeping the distance between thetwo Rfpulses constant), D can be obtained by fitting Eq. (1) to the observedintensities.

As mentioned earlier, the key feature of PFG diffusion experimentsis the fact that the transport of molecules is measured over a time �,which we are free to choose at our own will in the range of from a fewmilliseconds to several seconds. This means that the length scale overwhich we are measuring the molecular transport is in the micro-meter regime for low-molecular-weight liquids. When the molecules experi-ence some sort of boundary with regard to their diffusion during thetime �, the molecular displacement is lowered as compared to free diffu-sion, and the outcome of the experiment becomes drastically changed(2,11,12).

This situation applies to the case of restricted motion inside an emul-sion droplet. In this case, the molecular displacements cannot exceed thedroplet size, which indeed often is in the micrometer regime. Until recently,no analytical expressions that describe the echo decay in restricted geome-tries for arbitrary gradient pulses have been available. However,Barzykin(13) and Codd and Callaghan(14) have developed two approachesthat work for arbitrary gradient pulses. This is an important step and willundoubtedly lead to an increased applicability of the method. Because theapplication of these approaches are somewhat numerically cumbersome andPFG work performed up until now mainly rely on one of two approximateschemes, we will describe these schemes below.

In one of these, one considers gradient pulses, which are so narrowthat no transport during the pulse takes place. This has been termed theshort gradient pulse (SGP) (or narrow gradient pulse, NGP) limit. This caseleads to a very useful formalism whereby the echo attenuation can be writtenas (2)

E ,�, gð Þ ¼

ZZ� r0ð ÞP r0jr,�ð Þ exp i�g r� r0ð Þ½ � dr dr0 ð2Þ

where P(r0|r,�) is the propagator, which gives the probabilityof finding a spin at position r after a time � if it was originally atposition r0.

As discussed by Karger and Heink(15) and Callaghan (2), Eq. (2) canbe used to obtain the displacement profile, which is the probability for amolecule to be displaced dz during � in the direction of the field gradient,irrespective of its starting position. Modern NMR spectrometers are capableof producing gradient pulses with durations less than 1 ms and with


strengths around 10 T/m. Under these conditions, the SGP limit is oftenvalid. The challenge is to design E(, �, g) experiments that are sensitive to aparticular type of motion in a system of interest. The study topics includefree diffusion in the continuous phase in emulsions, restricted diffusioninside emulsion droplets, molecular exchange between emulsion dropletsin highly concentrated emulsions, or drug release from emulsion droplets.Some of those examples will be considered in this chapter.

For the case of free diffusion, P(r0|r, �) is a Gaussian function, and ifthis form is inserted in Eq. (2), Eq. (1) with the term ��/3 replaced with �is obtained, which is the SGP result for free diffusion. For cases other thanfree diffusion, alternate expressions for P(r0|r, �) have to be used. Tannerand Stejskal (16) solved the problem of reflecting planar boundaries; how-ever, the case of interest to us in the context of emulsion droplets, (i.e. thatof molecules confined to a spherical cavity of radius R) was presented byBalinov et al. (17). The result is

Eð,�, gÞ ¼9 �gR cosð�gRÞ � sinð�gRÞ½ �

2

�gRð Þ6

þ 6ð�gRÞ2X1

n¼0

j0nð�gRÞ� �2

�X

m

ð2nþ 1Þ�2nm�2nm � n2 � n

exp ��2nmD�

R2

� �1

�2nm � ð�gRÞ2� �2 ð3Þ

where jn(x) is the spherical Bessel function of the first kind and �nm is the mth

root of the equation j0nð�Þ ¼ 0. D is the bulk diffusion of the entrappedliquid, and the rest of the quantities are as defined earlier. The main pointto note about Eq. (3) is that the echo amplitudes do indeed depend on theradius, and thus the droplet radii can be obtained from the echo decay formolecules confined to reside in a sphere, provided that the conditions under-lying the SGP approximation are met.

The second approximation used is the so-called Gaussian phasedistribution. Originally introduced by Douglass and McCall (18), theapproach rests on the approximation that the phases accumulated bythe spins on account of the action of the field gradients are Gaussiandistributed. Within this approximation and for the case of a steadygradient, Neuman (19) derived the echo attenuation for moleculesconfined within a sphere, within a cylinder and between planes. Forspherical geometry, Murday and Cotts (20) derived the equation forpulsed field gradients in the Hahn echo experiment described earlier.


The result is

ln Eð,�, gÞ½ �

¼ �2�2g2

D

X1

m¼1

��4m

�2mR2 � 2

� 2�2þ exp ��2mDð�� Þ

� �� 2exp ��2mD

� �

�2mD

� �

�2exp ��2mD�

� �þ exp ��2mDð�þ Þ

� �

�2mDð4Þ

where �m is the mth root of the Bessel equation ð1=ð�RÞÞJ3=2ð�RÞ ¼ J5=2ð�RÞ.Again, D is the bulk diffusion coefficient of the entrapped liquid.

Thus, we have at our disposal two equations with which to interpretPFG data from emulsions in terms of droplet radii, neither of which areexact for all values of experimental and system parameters. As the condi-tions of the SGP regime are technically demanding to achieve, Eq. (4) (orlimiting forms of it) have been used in most cases to determine the dropletradii. A key question is then under what conditions Eq. (4) is valid. That itreduces to the exact result in the limit of R!1 is easy to show and alsoobvious from the fact that we are then approaching the case of free diffu-sion, in which the Gaussian phase approximation becomes exact.

Balinov et al. (17) performed accurate computer simulations aimed atfurther testing its applicability over a wide range of parameter values. Anexample is shown in Fig. 1. The conclusion reached in Ref. 17 was thatEq. (4) never deviates by more than 5% in predicting the echo attenuationfor typically used experimental parameters. Thus, it is a useful approxima-tion and we will use it in Section III.

As pointed out earlier, the NMR echo amplitude E depends on thedroplet’s radius, which can be estimated by measuring E at different dura-tions of and/or magnitudes g of the pulsed gradient. Typical echo attenua-tion profiles, generated by using Eq. (4), are presented in Fig. 2, whichdemonstrates the sensitivity of the NMR self-diffusion experiment to resolvemicrometer droplet sizes.

We end this section by noting that the above discussion applies to thesituation when the dispersed phase cannot exchange between droplets orbetween the droplets and the continuous medium (at least on the relevanttimescale). This is the case for most emulsions. However, for some concen-trated emulsions, this is not necessarily the case and the molecules of thedispersed phase may actually exchange between droplets during the char-acteristic time of the measurement. This leads to special effects, which willbe discussed in Section III.


III. STUDIES OF DIFFUSION AND FLOW IN EMULSIONS

A. Determination of Emulsion Droplet Radii by Means of the NMRDiffusometry Method

As pointed out earlier, the echo attenuation curve for the PFG experimentwhen applied to molecules entrapped in an emulsion droplet is a signature ofthe size of the emulsion droplet (cf. Fig. 2). As a consequence, droplet sizescan be determined by means of the PFG experiment.

The NMR sizing method, which was apparently first suggested byTanner in Ref. 10 has been applied to a number of different emulsionsranging from cheese and margarines to crude oil emulsions (21–27).

When applied to a real emulsion one has to consider the fact that theemulsion droplets in most cases are polydisperse in size. This effect can beaccounted for if the molecules confined to the droplets are in a slow-exchange situation, meaning that their lifetime in the droplet must belonger than �. For such a case, the echo attenuation is given by

Epoly ¼

R1

0

R3PðRÞEðRÞ dR

R1

0

R3PðRÞ dR

ð5Þ

where P(R) represents the droplet size distribution function and E(R)represents the echo attenuation according to Eq. (4) [or, within the SGPapproximation, Eq. (3)] for a given value of R.

Figure 1 Results of a simulation of the diffusion of water molecules inside an

emulsion droplet of radius R, given as the echo amplitude, versus the duration ofthe field gradient pulse. The ratio D�/R2 is 1. The dotted line is the prediction of the

Gaussian phase approximation [Eq. (4)], whereas the solid line is the prediction of

the SGP [Eq. (3)]. (Adapted from Ref. 17.)


The principle goal is to extract the size distribution function P(R) fromthe experimentally observed (E(,�, g)) (henceforth referred to as E) pro-files. Two approaches have been suggested. In the first, one assumes thevalidity of a model size distribution with a given analytical form (26,27),whereas in the second, the size distribution which best describes theobserved experimental E profiles is obtained without any assumption onthe form of the size distribution (28,29).

A frequently used form for the size distribution (26,27) is the log-normal function as defined in Eq. (6), as it appears to be a reasonabledescription of the droplet size distribution of many emulsions:

P Rð Þ ¼1

2Rffiffiffiffiffiffi2�

p exp �ln 2R� ln d0ð Þ

2

22

� �

ð6Þ

In addition, it has only two parameters, which makes it convenient formodeling purposes. In Eq. (6), d0 represents the diameter median and is ameasure of the width of the size distribution.

To illustrate the method and discuss its accuracy, we will use as anexample some results for margarines (low-calorie spreads) (21). This systemhighlights some of the definite advantages of using the NMR method todetermine emulsion droplet sizes, because other nonperturbing methodshardly exist for these systems.

Given in Fig. 3 is the echo decay for the water signal of a low-caloriespread containing 60% fat. These systems are w/o emulsions and as can be

Figure 2 The echo attenuation as a function of 2(��/3) for different radii of

emulsion droplets [according to Eq. (4)] with � ¼ 0.100 s, �g¼ 107 rad/m/s, and

D¼ 2� 10�9m2/s.


seen the water molecules do experience restricted diffusion (in the represen-tation of Fig. 3, the echo decay for free diffusion would be given by aGaussian function). Also given in Fig. 3 is the result of fitting Eqs. (4)–(6)to the data. As is evident, the fit is quite satisfactory and the parameters ofthe distribution function obtained are given in the figure caption. However,one might wonder how well determined these parameters are, given the factthat the equations describing the echo attenuation are quite complicated. Totest this matter further, Monte Carlo calculations were performed in Ref. 21.Thus, random errors were added to the echo attenuation and the leastsquares minimization was repeated 100 times, as described previously(30). A typical result of such a procedure is given in Fig. 4. As can beseen in Fig. 4 the parameters are reasonably well determined, with an uncer-tainty in R (and , data not shown) of about 15%.

In the second approach, Ambrosone et al. (28,29) have developed anumerical procedure based on a solution of the Fredholm integral equationto resolve the distribution function P(R) without prior assumptions of itsanalytical type. The method involves the selection of a generating functionfor the numerical solution, which may not be trivial in some cases. Themethod was tested by computer simulation of E for a hypothetical emulsionwith a bimodal droplet size distribution (31). Figure 5 shows the reconstruc-tion of the true droplet size distribution from the synthesized E data.

Creaming or sedimentation of emulsions with droplet sizes above 1 mmcause some experimental difficulty because of the change of the totalamount of spins in the NMR active volume of the sample tube during the

Figure 3 Echo intensities for the entrapped water in droplets formed in a low-

calorie spread containing 60% fat versus . The solid line corresponds to the

predictions of Eqs. (4)–(6). The results from the fit are d0¼ 0.82 mm and ¼ 0.72.



experiment. This can be accounted for by extra reference measurements withno gradient applied before and after each NMR scan at a particular value of or g. In addition, such reference measurements may provide informationon the creaming rate, which is a useful characteristic of emulsions. Creamingor sedimentation is not a problem in the study of most food emulsions (suchas low-calorie spreads) and highly concentrated emulsions.

Figure 5 Determination of the size distribution function in the case of a bimodal

distribution. The solid line represents the ‘‘true’’ volume fraction distribution

function. The dots represents its values evaluated from the generated NMR data.


Figure 4 A Monte Carlo error analysis of the data in Fig. 3. The value of the

parameter d0 in Eq. (6) is d0¼ 0.82 0.044mm (note that R0¼ d0/2 is plotted).



Emulsion droplet sizes in the range from 1 mm up to 50 mm can bemeasured with rather modest gradient strengths of about 1 T/m. Note thatthe size determination rests on determining the molecular motion of thedispersed phase, so the method cannot be applied to dispersed phaseswith low molecular mobility. In practice, oils with self-diffusion coefficientsabove 10�12m2/s are required for sizing of o/w emulsion. Of course, w/oemulsions with most conceivable continuous media can be sized.

Emulsion droplets below 1 mm in size can often be characterized by theBrownian motion of the droplet as such (exceptions are concentrated emul-sions or other emulsions where the droplets do not diffuse). This is theapproach taken in the study of microemulsion droplets, where the diffusionbehavior of the solubilized phase is characterized by the droplets (Gaussian)diffusion.

In conclusion, we summarize the main advantages of the NMR diffu-sion method as applied to emulsion droplet sizing. It is nonperturbing,requiring no sample manipulation (such as dilution with the continuousphase), and nondestructive, which means that the same sample may beinvestigated many times, which is important if one wants to study long-time stability or the effect of certain additives on the droplet size. It requiressmall amounts of sample (typically on the order of a few 100 mg).Moreover, the total NMR signal from the dispersed phase in emulsions isusually quite intense because of the large amount of spins. This fact allowsfor rapid measurements with a single scan per /g point.

B. Applications of NMR Techniques to Investigate Diffusion andFlow in Food Emulsions

Most of the applications of NMR diffusimetry on food emulsion arefocused on the droplet size measurement described earlier. Over the years,several groups have established (21,26,27) the usefulness of the methodwhen applied to margarine and low-calorie spreads (cf. Fig. 3 and 4) andthe method is now available on commercial benchtop NMR spectrometersfrom Bruker, Oxford Instruments, or Universal Systems. However, in theopen literature, there are relatively few articles that have been publisheddealing with droplet sizing of food emulsion. One contribution presentswater self-diffusion and fat self-diffusion in cheese (23). Callaghan et al.found no significant difference in the value of the diffusion coefficientsbetween Swiss and cheddar cheese. The water self-diffusion coefficientwas about one-sixth of the value of bulk water at 30�C and it was suggestedthat water diffusion is confined to the protein surface. The self-diffusion offat was reduced because of restrictions induced by the droplets.


Another study deals with the mobility of lipid contained in breadas a function of water content and temperature (32). The self-diffusioncoefficient obtained at 30�C is of the same order as the one measured forthe diffusion of lipids in bulk milk fat and was slightly affected by achange in water content (from 6% to 9%). The self-diffusion corre-sponds to the intrinsic motion within the lipid globule, and the distancesprobed were too small to allow the molecules to reach the dropletinterface.

Because certain NMR experiments are sensitive to motion, flow can bemeasured with appropriate experimental protocols. These techniques arereferred to as NMR velocimetry. The advantage of this technique is thata velocimetry profile or map can be obtained without assumptions of theuniformity of the shear rate. Because many food products exhibit complexnon-Newtonian rheological properties, NMR velocimetry across the annu-lar gap of a Couette cell allows for the visualization of nonuniform beha-vior. For example, the effect of the temperature on the rheology of butter fatand margarine has been studied (33). The results show the effect of thetemperature on the heterogeneity of the rheological status of the sample.Moreover, the transition from heterogeneous to homogeneous viscosity as afunction of the temperature was dependent on the emulsion compositionand structure.

IV. NMR DIFFUSION STUDIES OF CONCENTRATED EMULSIONS

The above discussion applies to the case where the molecules are confined tothe droplets on the timescale of the experiment. This is a reasonable assump-tion for many emulsions and it can in fact be tested by the NMR diffusionmethod by varying �. However, there are some interesting emulsion systemswhere this is not always the case. These are the so-called highly concentratedemulsions (often termed ‘‘high internal phase ratio’’ or ‘‘gel emulsions’’)(34,35), which may contain up to (and in some cases even more than)99% dispersed phase. Here, the droplets are separated by a liquid film,which may be very thin (on the order of 0.1 mm) and which may, in someinstances, be permeable to the dispersed phase.

The case when the persistent time of the dispersed phase in the dropletis of the same order of magnitude as � is particularly interesting. Underthese conditions, one may in some cases obtain one (or several) peak(s)in the plot of the echo amplitudes versus the quantity q (which is definedas q ¼ �g=2�). This is a surprising result at first sight, as we are accustomedto observe a monotonous decrease of the echo amplitude with q, but it isactually a manifestation of the fact that the diffusion is no longer Gaussian.


Such peaks can be rationalized within a formalism related to the one usedto treat diffraction effects in scattering methods (36), and the analysis ofthe data may yield important information regarding not only the sizeof the droplets but also the permeability of the dispersed phase throughthe thin films as well as the long-term diffusion behavior of the dispersedphase. In Fig. 6, we show the results, which display diffractionlike effectsin a concentrated emulsion system. The particular example pertains to aconcentrated three-component emulsion based on a nonionic surfactantwith the composition C12E4/C10H22/H2O(1 wt % NaCl) (3/7/90 wt%).The concentrated emulsion was made according to a protocol described inRef. 37.

The data in Fig. 6 are presented with the value of the quantity q(defined earlier) on the abscissa. This quantity has the dimension of inverselength and it is, in fact, related to the scattering vector used in describingscattering experiments. In fact, the inverse of the value of the position of thepeak can be related to the center-to-center distance of the droplets. In theexample given in Fig. 6, this value is roughly 1.8 mm, which is in roughagreement with twice the droplet radii as judged from microscope picturestaken of the emulsion.

In order to analyze the data in more detail, one needs access to atheory for diffusion in these interconnected systems. One such theory wasdeveloped by Callaghan and co-workers (12). It assumes that the SGP limit

Figure 6 Diffractionlike effects in a concentrated emulsion system based on

a nonionic surfactant with the composition C12E4/C10H22/H2O (1 wt% NaCl),

(3/7/90 wt%).


described earlier is valid and it is based on a number of underlying assump-tions of which pore equilibration is perhaps the most serious one. This latterassumption implies that an individual molecule in a droplet samples all ofthe positions in the interior of the droplet fully before it migrates to aneighboring droplet. The echo amplitude profile for such a case is givenby the product of a structure factor for the single pore and a function thatdepends on the motion of the molecules between the pores. The pore-hopping formalism takes as input the radius of the sphere, the long-timediffusion coefficient, the center-to-center distance between the droplets, and,finally, a spread in the center-to-center distance (to account for polydisper-sity of the droplets). The prediction of the pore-hopping theory for the datain Fig. 6 is included as a solid line obtained with theoretical pore-to-poredistance of 1.2 mm, spread of the pore size of 0.2 mm, and the long-timediffusion coefficient D¼ 9�10�11m2/s. The agreement is not quantitative(the difference most likely owing to problems in defining a relevant structurefactor for our system of polydisperse droplets), but the main features of theexperimental data are certainly reproduced. A value for the lifetime of watermolecule in a droplet (approximately 3ms) was obtained from the values ofthe above pore-hopping parameters.

A different starting point in the analysis of the data such as the one inFig. 6 is to make use of Brownian simulations (38). These are essentiallyexact within the specified model, although they do suffer from statisticaluncertainties. For the present case, one lets a particle perform a randomwalk in a sphere with a semipermeable boundary. With a given probability,the particle is allowed to leave the droplet, after which it starts to perform arandom walk in a neighboring droplet. That the approach yields peaks inthe echodecay curves can be seen in Fig. 7 (38). The simulation schemeyields essentially the same kind of information as the pore-hoppingtheory. Thus, one obtains the droplet size and the lifetime of a moleculein the droplet (or quantities related to this, such as the permeability of thefilm separating the droplet).

Clearly, data such as those presented in Fig. 6 can be used to studymany aspects of concentrated emulsions of which a few were exemplifiedearlier. It can also be used to study the evolution of the droplet sizewith time (recall that the method is nonperturbing) and also as afunction of changes in external parameters such as temperature,which is an important variable for the properties of nonionic surfactantfilms.

Finally, we note that concentrated emulsions are excellent model sys-tems in the development of the PFG methods as applied to the general classof porous systems, where the method has a great potential in providingrelevant and important information.


V. NMR DIFFUSION STUDIES OF MULTIPLE EMULSIONS

Multiple emulsions usually refer to series of complex two-phase systems thatresult from dispersing an emulsion into its dispersed phase. Such systems areoften referred to as water-in-oil-in-water (w/o/w) or oil-in-water-in-oil(o/w/o) emulsions, depending on the type of internal, intermediate, andcontinuous phase. Multiple emulsion were early recognized as promisingsystems for many industrial applications, such as in the process of immobi-lization of proteins in the inner aqueous phase (39) and as liquid membranesystems in extraction processes (40). The w/o/w emulsions have been dis-cussed in a number of technical applications [e.g., as prolonged drug deliv-ery systems (41–46), in the context of controlled-release formulations (47),and in pharmaceutical, cosmetic, and food applications (48)].

Multiple emulsions have a complex morphology and various impor-tant parameters for their preparation and characterization have beendescribed (41,49). Examples are the characteristics of the w/o globules inw/o/w systems, such as their size and volume fraction, water/oil ratioinside the w/o globules, and average number and size of water dropletsinside the w/o globules. The time dependence of those parameters is clo-sely related to the stability of multiple emulsions and their morphology.Other important features are transport properties of substances encapsu-lated into discrete droplets and the permeability of the layer separating theinternal from the external continuous phase. As was shown earlier, theNMR PFG method is a sensitive tool for studying structure and complex

Figure 7 Brownian dynamic simulations of the echo decay at various qR for

various q¼ �g/2� [R¼ 4mm, Pwall¼ 0.032 (the probability of a molecule to

penetrate the film), �¼ 100 ms]. The simulation scheme yields essentially the same

kind of information as the pore-hopping theory. (Adapted from Ref. 38.)


dynamic phenomena, therefore, it is a promising technique in the study ofmultiple emulsions.

An example of possible uses of the method is demonstrated in Fig. 8,where the echo signal from the water in a w/o emulsion (panel a) and fromthe resulting double emulsion obtained when the original w/o emulsions isemulsified in water (panel b) are displayed (5). Also given are the resultingsize distributions (panel c). The state of the water in multiple w/o/w emul-sions was examined by recoding the water proton spectra (50). A narrowsignal indicated water in a simple emulsion, whereas a broad signal indi-cated a multiple emulsion containing dispersed water.

Figure 8 The echo signal from the water in a w/o emulsion (panel a), and from the

resulting double emulsion obtained when the original w/o emulsions is emulsified in

water (panel b). Also given in panel c are the resulting size distributions (— is w/o

emulsion; - - - is w/o/w emulsion). Please note that the size distribution is not

normalized. (Adapted from Ref. 5.)


For studies of these complex systems, NMR is very well suited, as fewother methods exist that can determine such basic properties as the state ofwater and the size distribution of internal emulsion droplets. In addition, theNMR PFG method may convey information about the molecular transportfrom emulsion droplets, a quantity that is relevant in the context of releasemechanisms and rates from such emulsion carriers.

VI. DETERMINATION OF THE EMULSION COMPOSITION

Many industrial emulsion systems, such as cosmetic or pharmaceutical for-mulations, have well-defined compositions, whereas for other systems, thecomposition and nature of the ingredients may be unknown. An example ofthe latter is constituted by some food emulsions. The high sensitivity of theNMR experiment and ability to identify substances by their characteristicspectra may be used to quantify the emulsion composition. This methodrelies on the fact that the NMR spectra appears as a set of separated signalscorresponding to the nuclei of interest and where the separation is due to thevarying electronic environment within the molecule. The intensity of eachsignal, determined from the area under the signal, is proportional to thenumber of equivalent protons. Many emulsion systems are based on waterand hydrocarbon oils, which have well-resolved lines even in quite primitiveNMR spectrometers. This fact allows one to quantify the emulsioningredients of interest (such as oil, water surfactant, or additives) withoutthe need to separate the dispersed from the continuous phase. This quanti-tative characterization of emulsion systems may be particularly valuable forquality or process control, where an accurate and rapid analysis of theemulsion composition is a major requirement.

The determination of fat and water content in food emulsions withlow-field NMR spectrometer requires that the relaxation time (T2 or T1)from each fraction are different. This situation is often at hand when thefraction of nonfat dry matter is either sufficiently large to decrease the waterrelaxation time, as in seeds or powders (51), or sufficiently low, as in dres-sings. If the discrimination is not possible, the NMR measurement should beperformed before and after drying of the sample. A more sophisticatedtechnique has been proposed in Ref. 52, which is based on the differencebetween the water and the fat self-diffusion. By choosing appropriate para-meters of the NMR sequence, it was possible to determine both the waterand the fat content in the sample.

Other methods proposed are based on the chemical shift differencebetween the hydroxyl group from water and methyl groups from lipid.Tellier et al. (53) demonstrated that reliable NMR determinations of water


and fat in minced meat can also be obtained under flowing conditions, afinding which suggests future on-line applications. Small variations in thewater content (<4%) can be detected with a time response of less than 1min. However, the reliability of these measurements depends on the capacityof maintaining a constant temperature in the sample. Because high-resolu-tion NMR detects only liquid signals, temperature variations, which affectthe solid/liquid ratio of the fat, induce signal fluctuation. A high-field spec-trometer is not always necessary; fat and water content have also been deter-mined with a low-field high-resolution spectrometer (54). The accuracy of themethod was controlled on different meat products and cooked sausage. Themain advantage of the technique is that the measurement protocol did notdepend on the type of fat present. At low and medium resolution, anhydrousCuSO4 can be added in order to relax the large water resonance and enhancedetection of the much weaker oil resonance. This approach was used instudies of French salad dressing (55). These methods have been implementedon a magnetic resonance imaging (MRI) scanner to provide a tool to directlycontrol the water and fat content of a bottle of French-style dressing (56).The method is a proton density projection using a spin-echo imagingsequence with the phase encode gradient turned off. The accuracy of theMRI method is within 2%. A MRI method for simultaneous determina-tion of water and fat in cheese has also been proposed (57). The MRIsequence was a chemical shift selective sequence (CHESS). This sequenceprovides separately the fat- and water-enhanced images. Other MRIsequences have been also evaluated on dairy cream (58). The accuracy ofthese methods depends strongly on the main magnetic field strength, whichmay constitute a limitation for industrial applications. More recently, theMRI method has been developed for the control of the stability of the emul-sion. When the relaxation time for oil–water emulsion is known, the volumefraction of each component phase can be calculated using

1

T1obs¼�oilT1oil

þ�waterT1water

ð7Þ

where ’oil and ’water are the volume fractions of oil and water, respectively.A fast spatially localized technique was developed for T1 determination.Volume fractions along the profile of a 40% (v/v) oil–water emulsionwere calculated several times during creaming to establish the dynamics ofthe process (59). This method is based on the assumption that the relaxationtime of the fat and water phase was independent of the volume fraction.

The NMR characterization of emulsion composition may be quitevaluable also in fundamental studies of emulsion systems and their applica-tions. Various emulsion processing, such as creaming, solvent evaporation,


and extraction of substances by emulsions often require a quantitativeanalysis of the emulsion components, which can be performed with highprecision using NMR techniques.

VII. ESTIMATING THE CREAMING OR SEDIMENTATION RATE

It is difficult to measure an absolute value for the creaming (or sedimenta-tion) rate of an emulsion, as there is often a broad size distribution of theemulsion droplets. For an isolated droplet in a continuous medium, thecreaming/sedimentation rate is dependent on the difference in densitybetween the droplet and the continuous medium, the size of the droplet tothe second power, and the viscosity of the continuous medium. It is obviousthat different droplet sizes will give different creaming/sedimentation rates(assuming that there is a density difference between droplet and the contin-uous medium). We note two NMR methods that may be principally valu-able in obtaining information on emulsion creaming or sedimentation. Thefirst one is based on the quantitative analysis of the amount of dispersedphase. Emulsions containing large droplets gradually redistribute in a testtube and the creaming or sedimentation of the emulsion can be studied bydetermining the amount of droplet phase suspended in the emulsion at afixed position as a function of time. This could be done directly in the NMRtube or by analyzing the amount of dispersed phase in a sample withdrawnfrom a fixed position of the test tube at distinct time intervals (60).Additional centrifugation of the emulsion, followed by NMR comparisonof the composition of the lower and the upper fraction is a preferred methodfor more stable emulsions. The second method for estimating the sedimen-tation rate is based on flow measurements by NMR PFG. The method issensitive to flow rates in the range of micrometers per second along thedirection of the gradient of the magnetic field.

In many cases, the creaming or sedimentation occurs simultaneouslywith coalescence and is related to emulsion stability. In the next section, wewill briefly consider the assessment of emulsion shelf life by NMR.

VIII. DETERMINATION OF EMULSION SHELF LIFE AND EMULSIONSTABILITY

Traditionally, emulsion stability is characterized by evaluating the dropletsize distribution as a function of time and relating the results to variousformulation parameters. On the basis of such studies, the thermo-dynamic instability of conventional emulsions is understood and well


documented (61–63). As discussed earlier, the NMR PFG is able to yield theevolution of the droplet size distribution of the same sample over time with-out any destruction of the sample. The change in size distribution may thenbe interpreted in terms of the change in the average droplet size or the totaldroplet area. There are two mechanisms for the decrease in the total dropletarea. The first is the coalescence of two droplets involving the rupture of thefilm formed in the contact region of two neighboring droplets. The secondprocess is Ostwald ripening involving the exchange of the molecules of thedispersed phase through the continuous phase. For concentrated o/w emul-sions [at least the ones the present authors have investigated (64)], the per-meability across the thin liquid film between the droplets is so slow that thestability is given by the film rupture mechanism. Given in Fig. 9 is the totaldroplet area as a function of time for a concentrated emulsion consisting of98wt% heptane, water, and cetyl trimethylammonium bromide (CTAB). Ascan be seen, there is a rapid initial decrease in the area, which levels out afterabout 12 h. From the initial part of the curve, the film rupture rate may beobtained, and for the data in Fig. 9, the value is 4�10-5 s-1. At longer times,the emulsion becomes remarkable stable, and there is little or no furtherdecrease in the total droplet area during a period of 1 year.

In some concentrated emulsions, molecular exchange between theemulsion droplets occurs on a time-scale defined by the value of �. Thissituation is at hand if the dispersed phase crosses the film by somemechanism, the detailed nature of which need not concern us here.We are then dealing with a system with permeable barriers (on the relevant

Figure 9 Decrease of the relative droplet surface area (S/S0) with time for a highly

concentrated o/w emulsion containing 98wt% heptane and 0.4wt% CTAB as

emulsifier. S/S0 is calculated from the droplet size distribution as obtained by the

NMR self-diffusion technique. The initial slope corresponds to a film rupture rate of

J¼ 4� 10�5s�1. (Adapted from Ref. 64.)


timescale), and the system can now be regarded as belonging to the generalclass of porous systems. As was shown earlier, the interpretation of theNMR PFG signal in this case also gives a information on the dropletsize. Given in Fig. 10 is the droplet size determined at two different instances(65), demonstrating that the stability of a concentrated emulsion with timecan be conveniently followed by this method.

IX. STUDIES OF THE DISPERSED AND CONTINUOUS PHASE

A. Identification of the Dispersed Phase in Emulsions

Emulsions are formed by mixing two liquids, a process which creates dis-crete droplets in a continuous phase. During emulsification (e.g., by apply-ing mechanical agitation), both liquids tend to form droplets, resulting in acomplex mixture of o/w and w/o emulsions. Which of the components formthe continuous phase depends on the emulsifier used because one type of thedroplets is unstable and coalesce. Therefore, there is a need to identify thecontinuous phase in emulsion systems not only in the final emulsion systembut also at short times after emulsion formation or even during the emulsi-fication process. The NMR self-diffusion method may easily distinguish thecontinuous and dispersed phases based on the transport properties of itsmolecules. For example, molecules confined in the discrete droplets have anapparent diffusion coefficient, which is much lower than the correspondingvalue in the bulk phase. Molecules in the continuous phase, on the other

Figure 10 Normalized intensities versus q for one diffusion time (�¼ 50ms) at 2

(solid circles) and 8.5 (open circles) h after emulsion preparation. Parameters used in

the experiment were ¼ 3ms and a maximum gradient strength of 8.37T/m.



hand, have an apparent diffusion coefficient similar to the value in thecorresponding bulk phase and this fact can be used to identify the type ofcontinuous phase. This is particularly relevant in the case of emulsificationby the phase-inversion technique (66), where a single surfactant may formeither o/w or w/o emulsions depending of the formulation conditions. Inmany cases, the inversion of the emulsion from o/w to the w/o type (or fromw/o to o/w) is a required and important step of emulsion formation. Due toits sensitivity to the transport properties of the dispersed phase, the NMRself-diffusion method is a useful tool for studying the phase-inversionprocess.

B. Study of the Properties of the Continuous Phase

Extensive work has been carried out in order to establish a relationshipbetween emulsion properties and the properties of surfactant systems. Theclassical HLB (hydrophile–lipophile balance) concept is widely used inemulsion science to describe the balance of the hydrophilic and lipophilicproperties of a surfactant at oil–water interfaces. The HLB value determinesthe emulsion-inversion point (EIP) at which an emulsion changes from onetype to the other. This is of particular importance for nonionic surfactantsthat change their properties with changes in the temperature (66).Various NMR techniques have given significant contributions to the basicunderstanding of surfactant systems and some of those were reviewed inRef. 8. The usefulness of NMR in studying surfactant solutions lies in thedirect information they provide about the microstructure of microheteroge-neous systems (8,67–71). It is beyond the scope of this chapter to summarizethe use of NMR techniques to study surfactant systems, but we will presentsome representative examples related to emulsions.

In order to study the influence of the microstructure of the contin-uous phase on the stability of emulsions, two of the present authors inves-tigated the system sodium dodecyl sulfate (SDS)/glycerolmono(2-ethyl-hexyl)ether/decane/brine (3 wt% NaCl) (72). In this system, emulsions ofthe o/w or w/o type can be made depending on the ratio between surfac-tant and cosurfactant. The total amount of surfactant and cosurfactant iskept constant at 5wt%. The samples are made with equal weights of brineand decane and with a varying ratios between surfactant and cosurfactant.The emulsions in this system are made in two-phase areas of the phasediagram which for the o/w emulsions consists of an oil-rich phase and aphase of normal micelles and for the w/o emulsions consist of a water-richphase and a micellar phase of reversed micelles. The micellar phase is thecontinuous medium for both types of emulsion. In order to determine thestructure of the continuous medium, the emulsion samples were allowed to


cream (or sediment) and the clear continuous medium was separated fromeach sample. These solutions were then characterized by the NMR self-diffusion method and the diffusion coefficients of both the oil and thewater were determined. The result is shown in Fig. 11, where the reduceddiffusion coefficients (D/D0, where D is the actual diffusion coefficient andD0 is the diffusion coefficient of the neat liquid at the same temperature)for the oil and the water are plotted versus the ratio between cosurfactantand surfactant. For the o/w emulsion where SDS is the only surfactant,one finds that the continuous medium consists of small spherical micelles,that the water diffusion is fast [slightly lowered relative neat water due toobstruction effects (74)], and that the oil diffusion is low and correspondsto a hydrodynamic radius of the oil swollen micelle of about 50 A (accord-ing to Stokes law). When the cosurfactant is introduced and its amount isincreased, the size of the micelles is increased, as can be inferred from thediagram by the lowered value of the reduced diffusion coefficient of theoil. Close to the three-phase area, the continuous medium is bicontinuous,as the value of the reduced diffusion coefficient is almost the same forboth the oil and the water and is equal to approximately 50% of the valuefor the bulk liquids. When the amount of cosurfactant is increased further,one passes over to the w/o emulsion region, where the continuous mediumis bicontinuous near the three-phase area and then changes to closedreversed micellar aggregates, as can be seen from the reverse in order ofthe magnitudes of the values of oil and water diffusion coefficients.The hydrodynamic radius of the inverse micelles is about 70 A.

In another study (75), NMR self-diffusion measurements of the con-tinuous oil phase show that a stable highly concentrated w/o emulsion is

Figure 11 Microemulsion structure in the continuous phase studied by the

diffusion coefficients (D) divided by the diffusion coefficients (D0) for the neat liquid

versus the relative amount of SDS in the SDS surfactant mixture. �: decane

diffusion; f: water diffusion. (Adapted from Ref. 73.)


formed when the continuous phase is a reverse micellar solutions,whereas unstable emulsions were formed when the continuous phase is abicontinuous microemulsions.

C. 31P-NMR of Emulsion Components

The linewidths of 31P-NMR can be used to characterize the properties ofphospholipids. In emulsions, the linewidths are affected by the aqueousphase pH, size, or dispersion states of particles and methods of emulsifica-tion. The hydrophilic head-group motions of emulsified egg yolk phospha-tidyl-cholin (PC) and lyso PC are examined by 31P-NMR to evaluate theirphospholipid states and stability. The results suggest that the head-groupmotions of phospholipds are related to emulsion stability (76,77).

Many emulsion systems are stabilized by phospholipids that formvarious self-organized structures in the continuous phases. Examples arefat emulsions containing soy triacylglycerols and phospholipids that areused for intravenous feeding. Studies have shown that these emulsions con-tain emulsion droplets and excess phospholipids aggregated as vesicles (lipo-somes), which remain in the continuous phase upon separation of theemulsion droplets by ultracentrifugation. The lamellar structure of thevesicles in the supernatant was characterized by 31P-NMR (78) that distin-guishes lipids in the outer and inner lamellae. 31P-NMR was used (79) toconfirm that the resulting structures in lipid emulsions are emulsion dropletsrather than lipid bilayers. The composition of the fat particles of parenteralemulsions was obtained by 31P-NMR (80). Analysis of the data identifiedthe ratio of phospholipid/triacylglycerol in various fractions of emulsiondroplets separated by centrifugation. 31P-NMR showed (81) that approxi-mately 48mol% of the phospholipid emulsifier in a model intravenousemulsion forms particles smaller than 100 nm in diameter.

31P-NMR and 13C-NMR may be used to study the emulsifier proper-ties at the o/w interface. The analysis of T1 relaxation times of selected 13Cand 31P nuclei of b-casein in oil–water emulsions indicates (82) that con-formation and dynamics of the N-terminal part of b-casein are not stronglyaltered at the oil–water interface. A large part of the protein was found in arandom coil conformation with restricted motion with a relatively longinternal correlation time.

D. Molecular Organization from Relaxation Measurements

In addition to the NMR signal intensity, values of T1 and T2 have been usedto characterize food emulsions. Indeed, the relaxation time T2 for the waterphase is sensitive to biopolymer conformation changes through chemical


exchange mechanisms. Monitoring the water phase T2 relaxation providesinformation on the effects of emulsion formation, addition of surfactants,and temperature on the biopolymer structure at the fat droplet surface. Forexample, large differences in the modification of T2 induced by the forma-tion of emulsions have been observed when adding Na-casein or b-lactoglo-buline (83). Because of the initial conformation of the protein, the effect ofthe fat–protein interaction at the interface on the protein structure wasdifferent. The sensitivity of the relaxation time depends on the biopolymerratio between the amount fixed at the droplet surface and the amount dis-persed in the aqueous phase. If the biopolymer content in the water phase isin large excess, then the relaxation will not be sensitive to the interface. Thisis illustrated by the relaxation time behavior in cheese (84). In that case, thewater relaxation is related to the protein/water ratio and independent of thefat content, and the T2 relaxation time of the water phase is sensitive tothe molecular and microstructural organization of the protein gel (85). Themolecular structure of the protein gel induced during ripening could bemonitored with T2 relaxation measurements or T2 maps obtained fromMRI. With appropriate image analysis methods, the volume of the nonma-ture part and the mature part in cheese could be quantified (86) and relatedto sensory properties of the cheese (87,88).

The relaxation times of fat have also been studied. When the fat is inthe liquid state, the relaxation time T2 is influenced by the carbon chainlength and by the degree of unsaturation (89). As many fats in food are acomplex mixture of triglycerides, the NMR signal can be described by abroad distribution of relaxation times (84) and the interpretation of suchbehavior in food is difficult. In contrast to this situation, large changes inthe longitudinal relaxation time of solid fat according to the polymorphicstate have been observed in triglycerides (90). The longitudinal relaxationtime was 12 s at 25�C for the polymorphic b form and decreases to 0.6 swhen polymorphic � is formed. Measurement of T1 provides a method offollowing the time course of the polymorphic transformation upon, forexample, aging. Such variations have also been observed between fat inbulk and in emulsions for model food emulsions (83) or for real food pro-ducts such as ice cream mix (91). However, because of the complexity of fatcomposition, the interpretation in terms of polymorphism modification isdifficult.

X. DEGREE OF SOLIDIFICATION OF THE DISPERSED PHASE

Many emulsion-based formulations also contain solid particles. Typicalexamples in the area of food products are margarine and salad dressing.


In the food industry, the determination of the amount of solid fat is anessential part of process control. An example is the monitoring of the fathardening in margarine after its formation as a w/o emulsion. Close controlof the solid fat content (SFC) is needed to give the margarine its character-istic properties. The NMR methods used in the determination of the SFCwill be briefly described in Section X.A. We conclude the chapter by dis-cussing NMR methods in the context of emulsion stability.

A. Determination of the Solid Fat Content

The determination of the SFC content by pulsed NMR is based on the factthat the transverse magnetization of solid fat decays much faster than oil.The spin–spin relaxation time (T2) of solid fat is about 10 ms and that of oil isabout 100ms. The NMR signal, derived from the amplitude of the FID, ofpartially crystallized fat after a 90o radio-frequency (RF) pulse is schemati-cally shown in Fig. 12. The magnetization of the solid fat decays very fast.As a consequence, its contribution to the signal is far less than 0.1% of theinitial value after about 70 ms. The decrease of the liquid-oil signal at thismoment (70 ms) is smaller than 1% and, therefore, the signal intensity will bewell proportional to the number of protons in the liquid.

Two pulsed NMR methods exist for measuring the SFC of edible oils:the direct and indirect methods. The direct method determines the SFCfrom the signals corresponding to the total amount of solid and liquid fat(measured at time t� 0) and the amount of liquid fat (measured at long

Figure 12 The NMR signal of partially crystallized fat after a 90� RF pulse.


enough time). The indirect method, determines the SFC, Sind, by comparingthe signal from the liquid fraction Il with the signal Im from completelymelted fat. We have

Sind ¼cIm � Il

cIm� 100% ð8Þ

where the factor c corrects the signal for the temperature dependence ofboth the equilibrium magnetization and the Q-factor of the receiver coil.The correction factor c ¼ I0t=I0m is obtained by measuring the signals I0tand I0m of a reference liquid sample at both the measuring temperature andthe melting temperature, respectively. The indirect method is mainly used asa reference for the direct method when the signals from both the liquid andthe solid fat are processed. The indirect method is similar to the previouslyused continuous-wave (wide-line) technique that has some disadvantagescompared to the pulsed NMR approaches. For example, saturation condi-tions are needed to obtain a sufficiently large signal-to-noise ratio. Evenwith a well-chosen reference sample, the systematic error is in the range of1–2% (92). The wide-line analyses give only the narrow peak from the liquidfat, which should be compared to the signal of melted fat. This requires asecond measurement to be performed after at least 1 h and automation ofthe measurement is hardly possible.

The direct method solves these problems and measures the SFC basedon the processing of both the liquid and the solid signal (93). The measure-ment procedure may be fully automated and the percentage of solid isdisplayed immediately after the measurements. The measuring time is afew seconds and the SFC is determined with a standard deviation ofabout 0.4%. The signal can be obtained directly from the magnetizationdecay of the solid-fat protons and is equal to the signal immediately after the90� pulse; see Fig. 12. Due to the dead time of the receiver, it is not possibleto measure the initial signal height of solid and liquid, sþ l, but only a signals0 þ l after a certain time of about 10 ms after the 90� pulse. To determine theSFC, the ‘‘true’’ NMR signal from solid fat, s, may be obtained from themeasured signal, s0, from the solid fraction multiplied with the correctionfactor f which depend on the T2 of the solid fat protons. The solid fatfraction S can be expressed in terms of s0 (equal to the difference betweenthe observed and liquid signal according to Fig. 12) as

S ¼fs0

l þ fs0� 100% ð9Þ

The correction factor f¼ s/s0 can be determined from the measured s’and the solid fraction s obtained by the indirect method [Eq. (8)] using a


reference sample. The calibration should be performed once for a series ofsimilar samples.

The direct method is very fast and reproducible and the time forsample preparation is minimal. Only one NMR measurement is needed toobtain a SFC value.

Commercial spectrometers are available that converts the NMR signalto the percentage of SFC in fats and margarine. Examples of NMR spectro-meters that are suitable for the characterization of SFC are PC100,NMS100, the Minispec from Brucker, QP20þ from Oxford Instruments,and Maran Ultra from Universal Systems. Determination of SFC byNMR is a recognized international ISO standard (94).

In contrast to calorimetric methods, NMR allows the determinationof the SFC both under isothermal conditions and as a function of tem-perature. Consequently, many studies have been carried with the aim ofunderstanding the effect of the fat blend on the SFC, mainly focusing ondairy fat (95–97) and cacao butter (98). The NMR method was used alsofor process control of the crystallization behavior (99,100). For example,effects of minor components (101), effects of surfactants (102), effects ofinteresterification (103,104), and effects of protein–fat interactions (102)on the crystallization of dairy fat and complex fat mixtures have beenstudied.

The use of the direct method and the use of calibrated standards allowfor the determination of SFC expressed in mass fraction. Further assump-tions made are that the hydrogen nuclei contents and relaxation properties(T2) of the liquid and the standards are similar. For phase behavior studies,molar fractions, rather than mass fractions, of the solute in solvent arerequired. When all of the molecules in a sample blend are of similar mole-cular weight, the standard mass SFC provides a good approximation of themolar SFC. However, when the molecular weights of the blend solute andsolvent differ, mass SFC can differ greatly from molar SFC (105).

The direct or indirect methods are only valid for anhydrous fat onaccount of the complication posed by the presence of water protons con-tributing to the signal. The relaxation of water is somewhat faster than thatof oil, but not enough to distinguish the oil and water contributions to thesignal (95). Two strategies have been proposed to tackle this problem, one isbased on water suppression and the other subtracts the water phase signalfrom the total liquid signal. The water signal could be saturated by selectingshort repetition time between the pulses (106). However, accurate resultswith this method assume that the liquid-fat signal is not affected by the shortrepetition time. In Ref. 102, it was proposed to replace water by heavywater. This method is based on the assumption that the presence of D2Odoes not alter the behavior of the emulsion. Another method is to change


the relaxation time of water by adding paramagnetic ions (Cu2þ, Mn2þ), theaddition of which causes the relaxation time of water to become so smallthat the liquid-fat signal could be estimated. This method has been used inRef. 107 on cream, where the water phase signal was damped by addingMnCl2, whereas Walstra and van Beresteyn (108) used CuSO4. This methodrequires measurements of the water phase signal separately as well as knowl-edge of the exact fat fraction of the cream. Moreover, this method assumesthat the water and fat signals are additive. This latter assumption has beenvalidated in Ref. 109, where no interference between the signals were found.In order to study the influence of surfactant on ice cream mix emulsions,Barfod et al. performed two measurements: one on a water phase samplecontaining all water-soluble components and the other on the emulsion(110). Knowing the fat content of the mixture, the SFC could be calculatedfrom

%SFC ¼fs� 100

fsþ l � xlH2O

� � ð10Þ

l represents the liquid signal, lH2O is the liquid signal from the water phase,and x is the water fraction. This latter method was recently used to study theeffect fat crystallization on the behaviors of protein and lipids at a lipid–water interface (111).

As noted earlier, the determination of the SFC is based on thesampling of two points from the NMR signal: one after the dead time ofthe probe (11 ms) and the next one at 70 ms. This method assumes that therelaxation time of the solid part is temperature independent andindependent of the fat composition when using an external reference.This assumption is not always valid. For example, the precise analysis ofthe free-induction signal of milk fat at 5�C shows that an intermediatephase, characterized by a relaxation time T�

2 varying in the 50–250-msrange, could be detected (112). This intermediate phase represents about4% of the entire fat content and, consequently, it should be consideredwhen absolute values of SFC are required. An alternative is to analyzethe complete NMR signal. This approach provides a precise determinationof the relaxation time of the solid phase and the signal intensity is deducedwithout the need to correct for probe dead time. This method has beensuccessfully applied on fat blends (113), on model food emulsions(83,114), on cheese (84) and ice cream mix (91).

The NMR method for determining SFC has been implement on aMRI scanner (115). The procedure to quantify the rate of crystallization isessentially the same as for the indirect method in spectroscopic NMRexperiments. The method has been used to study the fat signal intensity


from pure chocolate bars as a function of the cooling rate (116). Thevariation in the signal intensity was interpreted in terms of variations inthe solid/liquid ratio of the cacao butter; hence, the existence of differentpolymorphic forms in the two samples was suggested. DifferentialScanning Calorimetry (DSC) experiments confirmed this interpretation.Applied on adipose tissues from pig carcasses, a strong linear relationshipwas found between the SFC and the fat composition and this allowed foran accurate mapping of the solid/liquid ratio in animal subcutaneousadipose tissues and in intermuscular fat tissues from different pieces(117). The primary information obtained from MRI studies of lipid crys-tallization is the crystal growth rate (115,118). Oil-weighted MR imageswere obtained using a T1-weighed spin-echo sequence. The water signalwas attenuated by using a small delay between the RF pulses. The crystal-lization dynamics was quantified from localized spectroscopy using a sti-mulated echo-pulse sequence. The method was then evaluated on modelsystem both in bulk and in emulsion system. The results show that the rateof lipid crystallization differs significantly from bulk to the oil–wateremulsion state and that difference is dependent on the nature of thetriglycerides (115). From NMR and MRI studies, the crystal growth hasbeen empirically modeled and revealed a significant trilaurin-locationdependence, confirming that the rates of the nucleation processes wereaffected both by the concentration of trilaurin and the heat transfer prop-erties of the sample and vessel (118).

B. Studies of the Emulsion Stability

It was demonstrated (119,120) that pulsed NMR may be used to measurethe extent of oil solidification during cooling of oil-in-water emulsions.Pulsed proton NMR can distinguish between the oil and aqueous phasesbecause of the large differences between the relaxation times of oil and waterprotons. On cooling, the NMR signal obtained for the aqueous phase isrelatively small compared to the oil phase during the entire cooling cycle.For example, the water signal was approximately 10 times lower than thewater signal for o/w emulsion containing approximately 50% hydrogenatedoil (121). When the lipid phase containing the lipophilic emulsifier is cooled,the NMR signal decreases as the sample solidifies. The NMR signals fromthe different types of protons of the emulsion system are approximatelyadditive (the sum of the aqueous and oil phases). In practice, the signalobtained for the supercooled emulsion is always larger than expectedfrom a mixture with a solid fat and water, indicating that emulsificationhas had an inhibiting effect on fat solidification. The dispersion of fat intosmall droplets suppresses the rate of solidification under supercooling due to


nucleation phenomena in confined emulsion droplets (122–124). The extentof solidification at supercooling is high for very large emulsion droplets andcorrelates with a low emulsion stability. Unstable emulsions show littlesupercooling, but those that are relatively stable to creaming and phaseseparation are resistant to oil solidification. The greater the degree of dis-persion, the slower the rate of phase separation. A correlation was madebetween the emulsion stability and the NMR signal from the emulsion whencompared to the NMR signal from the fat and water phase alone (121). Aparameter called ‘‘percent interaction’’ was derived from the NMR signal(120), the value of which correlated well with actual resistance of the emul-sion to creaming and phase separation during storage. For example, theNMR signals from an emulsion and its corresponding fat phases weredetermined for an emulsion containing 48% hydrogenated oil, 1% acety-lated monoglyceride (a 49% total fat phase), 1% Tween-20, and 50% water(a 51% aqueous phase) (121) as follows:

Fat phase signal 71.5� 0.49¼ 35.0Aqueous phase signal 7.51� 0.51¼ 3.8Expected emulsion signal 38.8Observed emulsion signal 41.5

The ‘‘interaction percentages’’ (Int%) were calculated using

Int% ¼s1 þ l þ w

sþ w� 1

� �

� 100% ð11Þ

where the observed emulsion signal (s1þ lþw) corresponding to the signalfrom the solidified fat fraction, s1, and not solidified fat, l, and water, w,was compared to the expected emulsion signal (sþw) corresponding to solidfat and water.

An emulsion is considered to be relatively stable if the observed NMRis more than 30% larger compared to the sum of the NMR signal fromcorresponding bulk water and oil (121). Pulsed NMR cooling curve mea-surements on emulsions offer an improved method for prediction of emul-sion stability. By using a flow-through cell in the NMR magnet, the rate andextent of phase separation was measured accurately (120). The method isuseful in selection of optimum types and levels of emulsifiers for each system(119). NMR measurements can also assist in optimizing surfactant blends toobtain stable emulsion. At the optimum emulsion formulation for a parti-cular oil, the ‘‘interaction percentage’’ will be at a maximum, which indi-cates higher emulsion stability.


XI. CONCLUSIONS

In the above sections, we have presented various applications of the NMRtechnique in the study of emulsions. NMR is a versatile spectroscopic tech-nique. This is also reflected in the span of questions pertaining to variousaspects of emulsions that can be addressed with the NMR technique. Thetopics covered there include the determination of droplet size distributions,aspects of emulsion stability, crystallization of fat, and determination of fatand water contents in food emulsions.

It seems quite clear that emulsions will become increasingly importantin more specialized applications in the future. In such applications, thedemands for accurate design of properties of the emulsion systems, process,and quality control will be a major challenge. As should be clear from theabove, NMR will become increasingly important in this regard. Controlleddelivery of drugs as well as the use in the administration of pesticides is twoexamples of emulsion technology that will require accurate and reproduciblecharacterization.

Due to its insensitivity to the physical nature of the sample and to itsnonperturbing character, it seems likely that different NMR methods willfind increasing use in studies of emulsions that are important in varioustypes of food.

ACKNOWLEDGMENT

The work was financially supported by the Swedish Board for Industrialand Technical Development (NUTEC).

REFERENCES

1. A. Kabalnov and H. Wennerstrom, Langmuir 12, 276–292 (1996).

2. P. T. Callaghan, Principles of Nuclear Magnetic Resonance Microscopy,

Clarendon Press, Oxford, 1991.

3. R. R. Ernst, G. Bodenhausen, and A. Wokaun, Principles of Nuclear Magnetic

Resonance in One and Two Dimensions, Oxford University Press, Oxford, 1987.

4. D. Canet, Nuclear Magnetic Resonance. Concepts and Methods, Wiley,

Chichester, 1996.

5. I. Lonnqvist, B. Hakansson, B. Balinov, and O. Soderman. J. Colloid Interf. Sci.

192, 66–73 (1996).

6. P. T. Callaghan, C. M. Trotter, and K. W. Jolley, J. Magn. Reson. 37, 247–259

(1980).

7. P. Stilbs, Prog. Nucl. Magn. Reson. Spectrosc. 19, 1–45 (1987).


8. O. Soderman and P. Stilbs, Prog. Nucl. Magn. Reson. Spectrosc. 26, 445–482

(1994).

9. E. O. Stejskal and J. E. Tanner, J. Chem. Phys. 42, 288–292 (1965).

10. J. E. Tanner, Thesis, University of Wisconsin–Madison, 1966.

11. P. T. Callaghan and A. Coy, in NMR Probes of Molecular Dynamics,

(P. Tycko, ed.), Kluwer Academic, Dordrecht, 1993.

12. P. T. Callaghan, A. Coy, T. P. J. Halpin, D. MacGowan, J. K. Packer, and

F. O. Zelaya. J. Chem. Phys. 97, 651–662 (1992).

13. A. V. Barzykin, Phys. Rev. B 58, 14,171–14,174 (1998).

14. S. L. Codd and P. T. Callaghan. J. Magn. Reson. 137, 358–372 (1999).

15. J. Karger and W. Heink, J. Magn. Reson. 51, 1–7 (1983).

16. J. E. Tanner and E. O. Stejskal, J. Chem. Phys. 49, 1768–1777 (1968).

17. B. Balinov, B. Jonsson, P. Linse, and O. Soderman, J. Magn. Reson. A 104,

17–25 (1993).

18. D. C. Douglass and D. W. McCall, J. Phys. Chem. 62, 1102 (1958).

19. C. H. Neuman, J. Chem. Phys. 60, 4508–4511 (1974).

20. J. S. Murday and R. M. Cotts, J. Chem. Phys. 48, 4938–4945 (1968).

21. B. Balinov, O. Soderman, and T. Warnheim, J. Am. Oil Chem. Soc. 71, 513–518

(1994).

22. B. Balinov, O. Urdahl, O. Soderman, and J. Sjoblom, Colloids Surfaces 82,

173–181 (1994).

23. P. T. Callaghan, K. W. Jolley, and R. Humphrey, J. Colloid Interf. Sci. 93,

521–529 (1983).

24. X. Li, J. C. Cox, and R. W. Flumerfelt, AIChE J. 38, 1671–1674 (1992).

25. I. Lonnqvist, A. Khan, and O. Soderman, J. Colloid Interf. Sci. 144, 401–411

(1991).

26. K. J. Packer and C. Rees, J. Colloid Interf. Sci. 40, 206–218 (1972).

27. J. C. van den Enden, D. Waddington, H. Van Aalst, C. G. Van Kralingen, and

K. J. Packer, J. Colloid Interf. Sci. 140, 105–113 (1990).

28. L. Ambrosone, A. Ceglie, G. Colafemmina, and G. Palazzo, J. Chem. Phys.

110, 797–804 (1999).

29. L. Ambrosone, A. Ceglie, G. Colafemmina, and G. Palazzo, J. Chem. Phys.

107, 10,756–10,763 (1997).

30. P. Stilbs and M. Moseley, J. Magn. Reson. 31, 55–61 (1978).

31. L. Ambrosone, G. Colafemmina, M. Giustini, G. Palazzo, and A. Ceglie, Prog.

Colloid Polym. Sci. 112, 86–88 (1999).

32. G. Roudaut, D. van Dusschoten, H. van As, M. A. Hemminga, and

M. Lemeste, J. Cereal Sci. 28, 147–155 (1998).

33. M. M. Britton and P. T. Callaghan, J. Texture Studies 31, 245–255 (2000).

34. K. J. Lissant, J. Colloid Interf. Sci. 22, 462–468 (1966).

35. H. M. Princen, J. Colloid Interf. Sci. 91, 160–175 (1983).

36. P. T. Callaghan, A. Coy, D. MacGowan, K. J. Packer, and F. O. Zelaya,

Nature (London) 351, 467–469 (1991).

37. R. Pons, I. Carrera, P. Erra, H. Kunieda, and C. Solans, Colloids Surfaces A 91,

259–266 (1994).


38. B. Balinov, P. Linse, and O. Soderman, J. Colloid Interf. Sci. 182, 539–548

(1996).

39. R. H. Engel, S. J. Riggi, and M. J. Fahrenbach, Nature 219, 856–857 (1968).

40. N. N. Li and A. L. Shrier, in Recent Developments in Separation Science (N. N.

Li, ed.), Chemical Rubber Co., Cleveland, OH, 1972.

41. S. Matsumoto, in Nonionic Surfactants (M. Schick, ed.), Marcel Dekker,

New York, 1987.

42. P. J. Taylor, C. L. Miller, T. M. Pollack, F. T. Perkins, and M. A. Westwood,

J. Hyg. (London) 67, 485–890 (1969).

43. L. A. Elson, B. C. Mitchlev, A. J. Collings, and R. Schneider, Rev. Eur. Etudes

Clin. Biol. 15, 87–90 (1970).

44. C. J. Benoy, L. A. Elson, and R. Schneider, Br. J. Pharmacol. 45, 135–136

(1972).

45. A. F. Brodin, D. R. Kavaliunas, and S. G. Frank, Acta Pharm. Suec. 15, 1–12

(1978).

46. D. Whitehill, Chem. Drug 213, 130–135 (1980).

47. R. K. Owusu, Z. Qinhong, and E. Dickinson, Food Hydrocolloids 6, 443–453

(1992).

48. E. Dickinson, J. Evison, and R. K. Owusu, Food Hydrocolloids 5, 481–485

(1991).

49. M. Frenkel, R. Shwartz, and N. Garti, J. Colloid Interf. Sci. 94, 174–178 (1983).

50. A. Rabaron, P. A. Rocha-Filho, C. Vaution, and M. Seiller, Study and

demonstration of multiple w/o/w emulsions by nuclear magnetic resonance, in

Fifth International Congress on the Technology of Pharmaceuticals, 1989.

51. R. E. Hester and D. E. C. Quine, J. Dairy Res. 44, 125–130 (1977).

52. H. Berg, Determination of fat and water content in meat with NMR, in Fifth

International Conference on Applications of Magnetic Resonance in Food

Science, 2000.

53. C. Tellier, M. Trierweiler, J. Lejot, and G. J. Martin, Analusis 18, 67–72 (1990).

54. J. P. Renou, A. Briguet, P. Gatellier, and J. Kopp, Int. J. Food Sci. Technol. 22,

169–172 (1987).

55. P. Fairbrother and D. N. Rutledge, Analusis 21, 113–117 (1993).

56. J. R. Heil, W. E. Perkins, and M. J. McCarthy, J. Food Sci. 55, 763 (1990).

57. R. Ruan, K. Chang, P. L. Chen, R. G. Fulcher, and E. D. Bastian, J. Dairy Sci.

80, 9–15 (1998).

58. S. L. Duce, M. H. G. Amin, M. A. Horsfield, M. Tyszka, and L. D. Hall, Int.

Dairy J. 5, 311–319 (1995).

59. R. J. Kauten, J. E. Maneval, and M. J. McCarthy, J. Food Sci. 56, 799–801

(1991).

60. M. Gangoda, B. M. Fung, and E. A. O’Rear, J. Colloid Interf. Sci. 116,

230–236 (1987).

61. C. Washington, Int. J. Pharm. 66, 1–21 (1990).

62. J. Boyd, C. Parkinson, and P. Sherman, J. Colloid Interf. Sci. 41, 359–370

(1972).

63. N. Weiner, Drug Dev. Ind. Pharm. 12, 933–951 (1986).


64. O. Soderman and B. Balinov, in Emulsions and Emulsion Stability (J. Sjoblom,

ed.), Marcel Dekker, New York, 1996.

65. B. Hakansson, R. Pons, and O. Soderman, Langmuir 15, 988–991 (1999).

66. L.Marszall, inNonionic Surfactants (M. Schick, ed.),MarcelDekker,NewYork,

1987.

67. B. Lindman, O. Soderman, and H. Wennerstrom, in Novel Techniques to

Investigate Surfactant Solutions (R. Zana, ed.), Marcel Dekker, New York,

1987.

68. B. Balinov, U. Olsson, and O. Soderman, J. Phys. Chem. 95, 5931–5936 (1991).

69. B. Lindman and P. Stilbs, in Microemulsion Systems (H.L. Rosano and M.

Clausse, eds.), Marcel Dekker, New York, 1987.

70. B. Lindman, K. Shinoda, U. Olsson, D. Andersen, G. Karlstrom, and H.

Wennerstrom, Colloids Surfaces 38, 205–224 (1989).

71. A. Khan, in Specialist Periodical Reports, Nuclear Magnetic Resonance (G. A.

Webb, ed.), The Royal Society of Chemistry, Cambridge, 1993.

72. K. Shinoda, H. Kunieda, T. Arai, and H. Saijo, J. Phys. Chem. 88, 5126–5129

(1984).

73. O. Soderman, I. Lonnqvist, and B. Balinov, in Emulsions—A Fundamental

and Practical Approach, (J. Sjoblom, ed.), Kluwer Academic, Dordrecht, 1992.

74. B. Jonsson, H. Wennerstrom, P. Nilsson, and P. Linse, Colloid Polym. Sci. 264,

77 (1986).

75. C. Solans, R. Pons, S. Zhu, H. T. Davis, D. F. Evans, K. Nakamura, and

H. Kunieda, Langmuir 9(6), 1479–1482 (1993).

76. K. Chiba and M. Tada, Nippon Nogei Kagaku Kaishi 62, 859–865 (1988).

77. K. Chiba and M. Tada, Agric. Biol. Chem. 53, 995–1001 (1989).

78. M. Rotenberg, M. Rubin, A. Bor, D. Meyuhas, Y. Talmon, and

D. Lichtenberg, Biochim. Biophys. Acta 1086, 265–272 (1991).

79. J. Drew, A. Liodakis, R. Chan, H. Du, M. Sadek, R. Brownlee, and W. H.

Sawyer, Biochem. Int. 22, 983–992 (1990).

80. J. Ferezou, T. L. Nguyen, C. Leray, T. Hajri, A. Frey, Y. Cabaret, J. Courtieu,

C. Lutton, and A. C. Bach, Biochim. Biophys. Acta 1213, 149–158 (1994).

81. K. Westesen and T. Wehler, J. Pharmaceut. Sci. 82, 1237–1244 (1993).

82. L. C. ter Beek, M. Ketelaars, D. C. McCain, P. E. Smulders, P. Walstra, and

M. A. Hemminga, Biophys. J. 70, 2396–2402 (1996).

83. F. Mariette, S. Querangal, M. Cambert, C. C. Lopez, and A. Riaublanc, in

Protein structural changes ar the O/W interface emulsion studied by NMR

relaxometry, Fifth International Conference on Applications of Magnetic

Resonance in Food Science, 2000.

84. B. Chaland, F. Mariette, P. Marchal, and J. de Certaines, J. Dairy Res. 67,

609–618 (2000).

85. B. Chaland, Thesis, University of Rennes, France, 1999.

86. B. Chaland, F. Mariette, P. Marchal, and J. de Certaines, Macro structural and

morphological changes during soft cheese ripening by MRI, in Fifth

International Conference on Applications of Magnetic Resonance in Food

Science, 2000.


87. F. Mariette, G. Collewet, P. Marchal, and J. M. Franconi, in Advances in

Magnetic Resonance in Food Science (P. Belton, B. Hills, and G. A. Webb,

eds.), The Royal Society of Chemistry, Cambridge, 1999.

88. F. Mariette, G. Collewet, P. Fortier, and J. M. Soulie, in Magnetic Resonance

in Food Science—A View to the Future (G. A. Webb, P. Belton, A. M. Gil, and

I. Delgadillo, eds.), The Royal Society of Chemistry, Cambridge, 2001.

89. D. J. LeBotlan and I. Helie, Analusis 22, 108–113 (1994).

90. I. J. Colquhoun and A. Grant, Solid State NMR spectrocopy of fats, in

Eurofood Chem V, 1989.

91. T. Lucas, S. Dominiawsyk, F. Mariette, and G. Alvarez, 1H NMR assessment

of water behaviour in ice cream, in Engineering Food, Eighth ICEF

International Conference Proceedings, 2000.

92. A. J. Haighton, K. van Putte, and L. F. Vermaas, J. Am. Oil Chem. Soc. 49,

153–156 (1972).

93. K. van Putte and J. van den Enden, J. Phys. E: J. Sci. Instrum. 6, 910–912

(1973).

94. ISO 8292: 1991, Animal and vegetable fats and oils: determination of solid

fat content, pulsed nuclear magnetic resonance method, International

Organization for Standardization (1991).

95. B. K. Mortensen, in Developments in Dairy Chemistry — 2. Lipids (P. F. Fox,

ed.), Applied Science, London, 1983.

96. N. M. Barfod, in Characterization of Food: Emerging Methods (A. G.

Gaonkar, ed.), Elsevier, New York, 1995.

97. M. L. Herrera, M. D. Gatti, and R. W. Hartel, Food Res. Int. 32, 289–298

(1999).

98. H. K. Leung, G. R. Anderson, and P. J. Norr, J. Food Sci. 50, 942–945 (1985).

99. K. E. Kaylegian and R. C. Lindsay, J. Dairy Sci. 75, 3307–3317 (1992).

100. B. Breitschuh and E. J. Windhab, J. Am. Oil Chem. Soc. 73, 1603–1610 (1996).

101. A. J. Wright, R. W. Hartel, S. S. Narine, and A. G. Marangoni, J. Am. Oil

Chem. Soc. 77, 463–475 (2000).

102. N. M. Barfod and N. Krog, J. Anal. Org. Chem. Soc. 64, 112–119 (1987).

103. A. G. Marangoni and D Rousseau, J. Am. Oil Chem. Soc. 75, 1265–1271

(1998).

104. D. Rousseau, K. Forestiere, A. R. Hill, and A. G. Marangoni, J. Am. Oil

Chem. Soc. 73, 963–972 (1996).

105. A. G. Marangoni, A. J. Wright, S. S. Narine, and R. W. Lencki, J. Am. Oil

Chem. Soc. 77, 565–567 (2000).

106. S. Shanbhag, M. P. Steinberg, and A. I. Nelson, J. Am. Oil Chem. Soc. 48, 11

(1971).

107. E. G. Samuelson and J. Vikelsoe, Milchwissenschaft 26, 621–625 (1971).

108. P. Walstra and E. C. H. Van Beresteyn, Netherlands Milk Dairy J. 29, 36–65

(1975).

109. M. A. J. S. Van Boekel, J. Am. Oil Chem. Soc. 51, 316–320 (1974).

110. N. M. Barfod, N. Krog, G. Larsen, and W. Buchheim, Fat Sci. Technol.

1, 24–29 (1991).


111. T. Sugimoto, T. Mori, J. Mano, T. A. Mutoh, Y. Shiinoki, and Y. Matsumura,

J. Am. Oil Chem. Soc. 78, 183–188 (2001).

112. D. Le Botlan and I. Helie, Anal. Chim. Acta 311, 217–223 (1995).

113. J. Van Duynhoven, I. Dubourg, G. J. Goudappel, and E. Roijers, J. Am. Oil

Chem. Soc. 79, 383–388 (2002).

114. D. Le Botlan, J. Wennington, and J. C. Cheftel, J. Colloid Interf. Sci. 226,

16–21 (2000).

115. C. Simoneau, M. J. McCarthy, R. J. Kauten, and J. B. German, J. Anal. Org.

Chem. Soc. 68, 481–487 (1991).

116. S. L. Duce, T. A. Carpenter, and L. D. Hall, Lebensm. Wiss. Technol. 23,

545–590 (1990).

117. A. Davenel, P. Marchal, A. Riaublanc, and G. Gandemer, in Advances in

Magnetic Resonance in Food Science (P. Belton, B. Hills, G. A. Webb, eds.),

The Royal Society of Chemistry, Cambridge, 1999.

118. S. Ozilgen, C. Simoneau, J. B. German, M. J. McCarthy, and D. S. Reid,

J. Sci. Food Agric. 61, 101–108 (1993).

119. J. Trumbetas, J. A. Fioriti, and R. J. Sims, J. Am. Oil Chem. Soc. 53(12),

722–726 (1976).

120. J. Trumbetas, J. A. Fioriti, and R. J. Sims, J. Am. Oil Chem. Soc. 55(2),

248–251 (1978).

121. J. L. Cavallo and D. L. Chang, Chem. Eng. Prog. 86, 54–59 (1990).

122. D. Turubull, J. Chem. Phys. 20, 411 (1952).

123. J. P. C. Cordiez, G. Grange, and B. Mutaftschiev, J. Colloid Interf. Sci. 85,

431–441 (1982).

124. D. Clausse, in Encyclopedia of Emulsion Technology, (P. Becher, ed.), Marcel



emulsiones de alimentos

Documents